Deconvoluting_cell-types.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Scasat to deconvolute cell-type\n",
    "\n",
    "__Prerequisites:__\n",
    "\n",
    "In order to use Scasat, you have to have the following tools installed in the machine where you are running Scasat\n",
    "1. [samtools](http://www.htslib.org)\n",
    "2. [bedtools](http://bedtools.readthedocs.io/en/latest/)\n",
    "3. [macs2](https://github.com/taoliu/MACS)\n",
    "\n",
    "If you are using jupyter from anaconda installation then you can install the tools with the following command\n",
    "* samtools: conda install -c bioconda samtools\n",
    "* bedtools: conda install -c bioconda bedtools \n",
    "* macs2: conda install -c bioconda macs2 (please note that macs2 only runs on python2.7 so you have to create a python2.7 environment if you are using python3\n",
    "\n",
    "This jupyter notebook introduces __S__ingle-__c__ell __A__TAC-seq __A__nalysis __T__ool (Scasat) to deconvolute three cell types from a complex cell mixture. Scasat analyses single cell ATAC-seq data to identify regulatory regions that separates cells from a complex population. This notebook is the tool mentioned in [___Murtuza Baker et. al.___]()\n",
    "\n",
    "__Experimental design:__\n",
    "\n",
    "Two classic oesophageal adenocarcinoma (OAC) cell lines, OE19, OE33 and one non-neoplastic HET1A cell line were mixed together to create the complex mixture of population. These three cell lines were mixed at equal proportion to create this mixture. Single cell ATAC-seq was then performed on those two replicates by loading on two separate C1 fluidigm chips using a $96$ well plate integrated fluidic circuit (IFC) and sequenced on an Illumina NextSeq. This experimental figure is shown in the figure below\n",
    "\n",
    "<img src=\"ExperimentalDesign.png\" alt=\"Experimental Desing\" style=\"width: 500px;\"/>\n",
    "\n",
    "\n",
    "__Objective:__\n",
    "\n",
    "* Establish single cell ATAC-seq pipeline.\n",
    "* Identify the three different cells from the complex mixture\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Import python packages\n",
    "\n",
    "We import the python packages."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import subprocess, os, csv, signal, rpy2\n",
    "import numpy as np\n",
    "import readline\n",
    "import rpy2.robjects\n",
    "import matplotlib.pyplot as plt\n",
    "import csv\n",
    "from operator import itemgetter\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This code toggles between showing/hiding the codes. The outputs and the discussions would always be visible."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<script>\n",
       "code_show=true;\n",
       "function code_toggle() {\n",
       " if (code_show){\n",
       " $('div.input').hide();\n",
       " } else {\n",
       " $('div.input').show();\n",
       " }\n",
       " code_show = !code_show\n",
       "}\n",
       "$( document ).ready(code_toggle);\n",
       "</script>\n",
       "<form action=\"javascript:code_toggle()\"><input type=\"submit\" value=\"Click here to toggle on/off the raw code.\"></form>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from IPython.display import HTML\n",
    "\n",
    "HTML('''<script>\n",
    "code_show=true;\n",
    "function code_toggle() {\n",
    " if (code_show){\n",
    " $('div.input').hide();\n",
    " } else {\n",
    " $('div.input').show();\n",
    " }\n",
    " code_show = !code_show\n",
    "}\n",
    "$( document ).ready(code_toggle);\n",
    "</script>\n",
    "<form action=\"javascript:code_toggle()\"><input type=\"submit\" value=\"Click here to toggle on/off the raw code.\"></form>''')\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "rpy2 package provides a simple and robust access to `R` from within `Python`. The following cells loads the package to run `R` code within `Python` notebook. To write and execute `R` code we will use the magic command `%%R` at the top of the cell. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "%load_ext rpy2.ipython"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Python Functions\n",
    "\n",
    "Below we define the python functions for Scasat. One of the advantage of Scasat is the use of both python and R programming language. We use python for file processing and R for doing the data analysis."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "def formatSummitBed(summitFile, fileToSave):\n",
    "#fileToSave = 'location_toSave/AllFiles_Filtered_summits_shifted.bed' \n",
    "#peakFile = 'location_PeakFile/AllFiles_Filtered_summits.bed'\n",
    "\n",
    "    with open(summitFile,'r') as f:\n",
    "        peaks=[x.strip().split('\\t') for x in f]\n",
    "    for i in range(len(peaks)):\n",
    "        peaks[i][1] = str(int(peaks[i][1]) - 250)\n",
    "        peaks[i][2] = str(int(peaks[i][2]) + 250)\n",
    "\n",
    "    peaks.sort(key=lambda x: float(x[4]), reverse=True)\n",
    "\n",
    "    with open(fileToSave, \"w\") as f:\n",
    "        writer = csv.writer(f,delimiter='\\t')\n",
    "        writer.writerows(peaks)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "`formatSummitBed()`: \n",
    "> This functions formats the summit file by adding 500 bp distance from the summit position.\n",
    "\n",
    "Arguments:\n",
    "> `summitFile` : The name of the summit file. This has to be the file that is generated from the aggreagated `BAM` file<br>\n",
    "> `fileToSave`: The name of the file where the save the peak file with the 500 bp peaks.\n",
    "\n",
    "Value:\n",
    "> File with 500 bp peaks."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "def getMergedBam(BAMFolder, BAMFilePattern):\n",
    "    \n",
    "    import pandas as pd\n",
    "    \n",
    "        \n",
    "    mergeFiles = ''\n",
    "    for root, folder, files in os.walk(BAMFolder):\n",
    "            files = [os.path.join(root, f) for f in files if (f.endswith(BAMFilePattern))]\n",
    "            for f in files:\n",
    "                mergeFiles = mergeFiles +  f + ' '\n",
    "    \n",
    "        \n",
    "    mergeCommand = 'samtools merge ' + BAMFolder + 'AllFiles_Filtered_Merged_nodup.bam' + ' ' + mergeFiles    \n",
    "    subprocess.call(mergeCommand, shell=True)   \n",
    "    \n",
    "    indexMergeCmd = 'samtools index ' + BAMFolder + 'AllFiles_Filtered_Merged_nodup.bam'\n",
    "    subprocess.call(indexMergeCmd, shell=True)\n",
    "    \n",
    "    cleanCommand = 'samtools view -b ' + BAMFolder + 'AllFiles_Filtered_Merged_nodup.bam' + ' ' \\\n",
    "                    + 'chr1 chr2 chr3 chr4 chr5 chr6 chr7 chr8 chr9 chr10 chr11 chr12 chr13 chr14' \\\n",
    "                    + ' chr15 chr16 chr17 chr18 chr19 chr20 chr21 chr22 chrX  > ' \\\n",
    "                    + BAMFolder + 'AllFiles_Filtered_Merged_nodup_cleaned.bam'\n",
    "    subprocess.call(cleanCommand, shell=True)            \n",
    "   \n",
    "    sortMergeCmd = 'samtools sort ' + BAMFolder + 'AllFiles_Filtered_Merged_nodup_cleaned.bam' \\\n",
    "                    + '>' + BAMFolder + 'AllFiles_Filtered_Merged_nodup_cleaned_sorted.bam'        \n",
    "    subprocess.call(sortMergeCmd, shell=True)\n",
    "    \n",
    "    indexMergeCmd = 'samtools index ' + BAMFolder + 'AllFiles_Filtered_Merged_nodup_cleaned_sorted.bam'\n",
    "    subprocess.call(indexMergeCmd, shell=True) "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "`getMergedBam()`:\n",
    "> merges all the individual `BAM` files for each of the cells and creates `AllFiles_Filtered_Merged_nodup.bam` file. This file is then sorted and indexed.\n",
    "\n",
    "Arguements:\n",
    "> `BAMFolder`: Location of the folder where all the BAM files for each of the single-cells are. <br>\n",
    "> `BAMFilePattern` : The end pattern of the `BAM` files to be merged.\n",
    "\n",
    "Value:\n",
    "> The merged `BAM` file `AllFiles_Filtered_Merged_nodup_sorted.bam`. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "def getAggregatedPeak(BAMFolder, BAMFilePattern):\n",
    "    getMergedBam(BAMFolder, BAMFilePattern)\n",
    "    \n",
    "    MergedMacs2Cmd = 'source activate python2; /home/mqbsxsm2/anaconda3/envs/python2/bin/macs2 callpeak -t ' + BAMFolder + 'AllFiles_Filtered_Merged_nodup_cleaned_sorted.bam' +' -n ' \\\n",
    "                    + BAMFolder + 'AllFiles_Filtered' +' -q 0.2 -g hs -f BAMPE --nomodel --nolambda \\\n",
    "                    -B --call-summits --keep-dup all; source deactivate python2'\n",
    "    subprocess.call(MergedMacs2Cmd, shell=True)\n",
    "    \n",
    "    narrowPeakFile = BAMFolder + 'AllFiles_Filtered_peaks.narrowPeak'\n",
    "    mergedPeakFile = BAMFolder + 'mergedPeaks.bed'\n",
    "    tmpPeakFile = BAMFolder + 'tmpPeak.txt'\n",
    "    \n",
    "    \n",
    "    with open(narrowPeakFile,'r') as f:\n",
    "        peaks=[x.strip().split('\\t') for x in f]\n",
    "    peaks.sort(key=lambda x: float(x[8]), reverse=True)\n",
    "    \n",
    "    \n",
    "    peaksToWrite = list()\n",
    "    for i in range(len(peaks)):\n",
    "        peaksToWrite.insert(i, peaks[i][0:4])\n",
    "        peaksToWrite[i].append(peaks[i][8])      \n",
    "    \n",
    "    \n",
    "    with open(mergedPeakFile, \"w\") as f:\n",
    "        writer = csv.writer(f,delimiter='\\t')        \n",
    "        writer.writerows(peaksToWrite)        \n",
    "        \n",
    "    sortCmd = 'sort -k1,1 -k2,2n ' + mergedPeakFile + ' > ' + tmpPeakFile  \n",
    "    subprocess.call(sortCmd, shell=True)\n",
    "    \n",
    "    bedToolsCmd ='bedtools merge -i ' + tmpPeakFile + ' -c 5 -o last > ' + mergedPeakFile\n",
    "    subprocess.call(bedToolsCmd, shell=True)\n",
    "    \n",
    "    sortCmd = 'sort -nrk4 ' + mergedPeakFile + ' > ' + tmpPeakFile  \n",
    "    subprocess.call(sortCmd, shell=True)    \n",
    "    \n",
    "    copyCmd = 'cat ' + tmpPeakFile + ' > ' + mergedPeakFile\n",
    "    subprocess.call(copyCmd, shell=True)    "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "`getAggregatedPeak()`:\n",
    "> This function calls `MACS2` on the aggregated `BAM` file and generates the peaks. If the user is following the merging BAM option then this is the function they need to call. Then the `getMergedBam()` is called within this function\n",
    "\n",
    "Arguements:\n",
    "> `BAMFolder`: Location of the folder where all the BAM files for each of the single-cells are. <br>\n",
    "> `BAMFilePattern` : The end pattern of the `BAM` files to be merged.\n",
    "\n",
    "Value:\n",
    "> Peak files from MACS2 call."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "def getAggregatedPeakSE(BAMFolder, BAMFilePattern):\n",
    "    getMergedBam(BAMFolder, BAMFilePattern)\n",
    "    \n",
    "    MergedMacs2Cmd = 'source activate python2; /home/mqbsxsm2/anaconda3/envs/python2/bin/macs2 callpeak -t ' + BAMFolder + 'AllFiles_Filtered_Merged_nodup_cleaned_sorted.bam' +' -n ' \\\n",
    "                    + BAMFolder + 'AllFiles_Filtered' +' -q 0.2 -g hs -f BAM --nomodel --nolambda \\\n",
    "                    -B --call-summits --keep-dup all; source deactivate python2'\n",
    "    subprocess.call(MergedMacs2Cmd, shell=True)\n",
    "    \n",
    "    narrowPeakFile = BAMFolder + 'AllFiles_Filtered_peaks.narrowPeak'\n",
    "    mergedPeakFile = BAMFolder + 'mergedPeaks.bed'\n",
    "    tmpPeakFile = BAMFolder + 'tmpPeak.txt'\n",
    "    \n",
    "    \n",
    "    with open(narrowPeakFile,'r') as f:\n",
    "        peaks=[x.strip().split('\\t') for x in f]\n",
    "    peaks.sort(key=lambda x: float(x[8]), reverse=True)\n",
    "    \n",
    "    \n",
    "    peaksToWrite = list()\n",
    "    for i in range(len(peaks)):\n",
    "        peaksToWrite.insert(i, peaks[i][0:4])\n",
    "        peaksToWrite[i].append(peaks[i][8])      \n",
    "    \n",
    "    \n",
    "    with open(mergedPeakFile, \"w\") as f:\n",
    "        writer = csv.writer(f,delimiter='\\t')        \n",
    "        writer.writerows(peaksToWrite)        \n",
    "        \n",
    "    sortCmd = 'sort -k1,1 -k2,2n ' + mergedPeakFile + ' > ' + tmpPeakFile  \n",
    "    subprocess.call(sortCmd, shell=True)\n",
    "    \n",
    "    bedToolsCmd ='bedtools merge -i ' + tmpPeakFile + ' -c 5 -o last > ' + mergedPeakFile\n",
    "    subprocess.call(bedToolsCmd, shell=True)\n",
    "    \n",
    "    sortCmd = 'sort -nrk4 ' + mergedPeakFile + ' > ' + tmpPeakFile  \n",
    "    subprocess.call(sortCmd, shell=True)    \n",
    "    \n",
    "    copyCmd = 'cat ' + tmpPeakFile + ' > ' + mergedPeakFile\n",
    "    subprocess.call(copyCmd, shell=True)  "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "`getAggregatedPeak()`:\n",
    "> This function calls `MACS2` on the aggregated `BAM` file mapped as single end reads and generates the peaks. If the user is following the merging BAM option then this is the function they need to call. Then the `getMergedBam()` is called within this function\n",
    "\n",
    "Arguements:\n",
    "> `BAMFolder`: Location of the folder where all the BAM files for each of the single-cells are. <br>\n",
    "> `BAMFilePattern` : The end pattern of the `BAM` files to be merged.\n",
    "\n",
    "Value:\n",
    "> Peak files from MACS2 call."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "------------------\n",
    "## R functions"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Loading R packages (if not available it will install it before loading)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/mqbsxsm2/anaconda3/lib/python3.6/site-packages/rpy2/rinterface/__init__.py:145: RRuntimeWarning: Loading required package: pacman\n",
      "\n",
      "  warnings.warn(x, RRuntimeWarning)\n"
     ]
    }
   ],
   "source": [
    "%%R\n",
    "if (!require(\"pacman\")) install.packages(\"pacman\")\n",
    "pacman::p_load(ggplot2, ComplexHeatmap, Rtsne, prabclus, GGally, rtracklayer, corrplot, \n",
    "               cluster, EnsDb.Hsapiens.v86, ChIPpeakAnno, org.Hs.eg.db, rGREAT)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### R functions for the Scasat\n",
    "\n",
    "Now I define all the R functions. These functions will be used for most of the downstream analysis. A short explanation of each function along with its input and output are also given below of each function. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "%%R\n",
    "peakAccessibility <- function(mergedPeakFile, peakFolder, peakFilePattern, outputFolder, outputPeakFileName){\n",
    "    if(length(list.files(path=peakFolder, pattern = paste0(peakFilePattern,'.bed'), full.names=TRUE))>0){\n",
    "        files = list.files(path=peakFolder, pattern = paste0(peakFilePattern,'.bed'), full.names=TRUE)\n",
    "        unlink(files)\n",
    "    }\n",
    "    formatingToBed(peakFolder, peakFilePattern, outputFolder)\n",
    "    peakFilePattern = paste0(peakFilePattern,'.bed')\n",
    "    query = import(mergedPeakFile)\n",
    "    queryDF <- data.frame(query)\n",
    "    #queryDF <- queryDF[order(queryDF$name, decreasing=TRUE),]    \n",
    "    \n",
    "    totalOverlap <- data.frame(seqnames = queryDF$seqnames, start = queryDF$start, end = queryDF$end)    \n",
    "    files <- list.files(path=peakFolder, pattern = peakFilePattern, full.names=TRUE)\n",
    "    cellName <- list.files(path=peakFolder, pattern = peakFilePattern)\n",
    "    for (i in 1:length(files)){\n",
    "        subject = import(files[i])\n",
    "        hits = findOverlaps(query, subject)\n",
    "        hitsDF <- data.frame(hits)\n",
    "        cellName[i] <- gsub(peakFilePattern, '', cellName[i])\n",
    "        totalOverlap[hitsDF$queryHits, cellName[i]] <- 1\n",
    "        totalOverlap[-hitsDF$queryHits, cellName[i]] <- 0\n",
    "    }\n",
    "    outputFile = paste0(outputFolder,outputPeakFileName)\n",
    "    write.table(totalOverlap, outputFile, row.names=FALSE)    \n",
    "}    \n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "`peakAccesibility()`:\n",
    "> This function calculates the accessible peaks for each of the cells. It looks at each of the peaks listed in _Aggregated-Peak_ table and tries to find out whether there is any overlap between these peaks and the peaks of each cell. If there is then it puts a 1 agains that cells for that peak. Otherwise it puts 0. Calling these function for each of the cell completes the _Aggregated-Peak_ for which rows represnt all the peaks and the column represents cells.\n",
    "\n",
    "Arguements:\n",
    "> `mergedPeakFile`: This is the name and location of `mergedPeaks.bed`.<br>\n",
    "> `peakFolder`: This is the location of the folder where all the peaks for each of the single-cells are located. <br>\n",
    "> `peakFilePattern` : The end pattern of the `peak` files of each single cells found in the `peakFolder`. Currently the only two accepted patterns are `_peaks.narrowPeak` and `_summits_shifted.bed` <br>\n",
    "> `outputFolder`: Folder where all the outputs will be saved\n",
    "> `outputPeakFileName`: Name of the file where the peak accessibility table, _Aggregated-Peak_ would be saved. This file will contain all the accessibility information for all the cells.\n",
    "\n",
    "\n",
    "Value:\n",
    "> Peak accessibile table saved according to the name set in `outputPeakFileName`. The table will have the results in 1 or 0 expression peak opening or closing respectively."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "%%R\n",
    "formatingToBed <- function(peakFolder, peakFilePattern, outputFolder){\n",
    "    files <- list.files(path=peakFolder, pattern = peakFilePattern, full.names=TRUE)\n",
    "    for (i in 1:length(files)){\n",
    "        narrowPeak <- read.csv(files[i], header=FALSE, sep='\\t')\n",
    "        write.table(narrowPeak[,1:4], paste0(files[i],'.bed'), \n",
    "                    row.names=FALSE, sep='\\t', col.names=FALSE, quote=FALSE)\n",
    "    }\n",
    "}"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "`formatingToBed()`:\n",
    "> This function formats the bed file suitable for processing by Scasat. It adds a `.bed` extension at the end.\n",
    "\n",
    "Arguements:\n",
    "> `mergedPeakFile`: This is the name and location of `mergedPeaks.bed`.<br>\n",
    "> `peakFolder`: This is the location of the folder where all the peaks for each of the single-cells are located. <br>\n",
    "> `peakFilePattern` : The end pattern of the `peak` files of each single cells found in the `peakFolder`. Currently the only two accepted patterns are `_peaks.narrowPeak` and `_summits_shifted.bed` <br>\n",
    "> `outputFolder`: Folder where all the outputs will be saved\n",
    "\n",
    "Value:\n",
    "> Formats the bed files with adding a `.bed` extesion."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "%%R\n",
    "calcEntropy <- function(data){\n",
    "\tfreqs <- table(data)/sum(table(data))\n",
    "\treturn(-sum(freqs * log(freqs)))\n",
    "}"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "`calcEntropy()`:\n",
    "> Calcuates entropy is calculated as\n",
    "$$\\sum_{i=1}^c -p_i log_2 p_i$$\n",
    "\n",
    "\n",
    "`Usage`:\n",
    "> `calcEntropy(count_data)`\n",
    "\n",
    "Arguements:\n",
    "> `count_data`: The peak information converted into binary. \n",
    "\n",
    "Value:\n",
    "> Returns entropy of the count data."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "%%R\n",
    "getDiffAccessInformationGain <- function(cdBinary, groups=NULL){\n",
    "\n",
    "    if (length(levels(groups)) != 2) {\n",
    "        stop(paste(\"ERROR: wrong number of levels in the grouping factor (\", \n",
    "            paste(levels(groups), collapse = \" \"), \"), but must be two.\", \n",
    "            sep = \"\"))\n",
    "    }\n",
    "    if (is.null(groups)){\n",
    "        stop(\"ERROR: groups factor is not provided\")\n",
    "    }\n",
    "    \n",
    "    \n",
    "    SingleCell.Group1.CellNames <- names(groups[groups==levels(groups)[1]])\n",
    "    SingleCell.Group2.CellNames <- names(groups[groups==levels(groups)[2]])\n",
    "    \n",
    "    SingleCell.Group1.Binary <- cdBinary[,SingleCell.Group1.CellNames]\n",
    "    SingleCell.Group2.Binary <- cdBinary[,SingleCell.Group2.CellNames]\n",
    "    \n",
    "    CellType <- data.frame(CellType=gsub('_.*','',names(groups)))\n",
    "    \n",
    "         \n",
    "    SingleCell.Group1VsGroup2 <- cbind(SingleCell.Group1.Binary,SingleCell.Group2.Binary)\n",
    "    SingleCell.Group1VsGroup2 <- t(SingleCell.Group1VsGroup2)\n",
    "    #SingleCell.Group1VsGroup2 <- cbind(SingleCell.Group1VsGroup2,CellType)   \n",
    "\n",
    "    \n",
    "    dataDim = (dim(SingleCell.Group1VsGroup2)[2])\n",
    "    #dataDim = 5\n",
    "    \n",
    "    information.gain = vector(mode=\"numeric\", length=(ncol(SingleCell.Group1VsGroup2)-1))\n",
    "    SingleCell.Group1VsGroup2.res <- data.frame(Chr = cdBinary[1:dataDim,1],\n",
    "                                           Start = cdBinary[1:dataDim,2],\n",
    "                                           end = cdBinary[1:dataDim,3])        \n",
    "    \n",
    "    \n",
    "    for(i in 1:dataDim)\n",
    "    {\n",
    "        jointData = cbind(SingleCell.Group1VsGroup2[,i], CellType)\n",
    "        \n",
    "        gain <- calcEntropy(jointData[,2]) - \n",
    "                (sum(table(jointData)[c(1,3)])/length(jointData[,2]))*calcEntropy(jointData[jointData[,1]==0,]) - \n",
    "                (sum(table(jointData)[c(2,4)])/length(jointData[,2]))*calcEntropy(jointData[jointData[,1]==1,])\n",
    "        SingleCell.Group1VsGroup2.res$information.gain[i] <- gain        \n",
    "    }\n",
    "    \n",
    "        \n",
    "    SingleCell.Group1.rawAvg <- rowSums(SingleCell.Group1.Binary==1)/dim(SingleCell.Group1.Binary)[2]\n",
    "    SingleCell.Group2.rawAvg <- rowSums(SingleCell.Group2.Binary==1)/dim(SingleCell.Group2.Binary)[2]\n",
    "    \n",
    "    SingleCell.Group1VsGroup2.res$rawMeanGroup1 <- SingleCell.Group1.rawAvg\n",
    "    SingleCell.Group1VsGroup2.res$rawMeanGroup2 <- SingleCell.Group2.rawAvg\n",
    "    SingleCell.Group1VsGroup2.res$log2FoldChange <- log2(SingleCell.Group1VsGroup2.res$rawMeanGroup1/\n",
    "                                                       SingleCell.Group1VsGroup2.res$rawMeanGroup2)\n",
    "    \n",
    "\n",
    "    #write.csv(SingleCell.Group1VsGroup2[order(SingleCell.Group1VsGroup2$information.gain, decreasing=TRUE),], \n",
    "    #          paste0(levels(groups)[1],'_vs_',levels(groups)[2],'InformationGain.csv'), row.names=FALSE)\n",
    "    \n",
    "    SingleCell.Group1VsGroup2.res <- SingleCell.Group1VsGroup2.res[!is.nan(SingleCell.Group1VsGroup2.res$information.gain) &\n",
    "                                                                  !is.na(SingleCell.Group1VsGroup2.res$information.gain),]\n",
    "    \n",
    "    return(SingleCell.Group1VsGroup2.res)    \n",
    "}"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "`getDiffAccessInformationGain()`:\n",
    "> Caculates differential accessibility between two groups of cells using entropy and information gain. For each peak the accessibility is encoded as a binary variable. For each peak we calculate the \"Information gain\" by having the peak divided into two groups of cells for which we are trying to find out the differential accessibilty. The Information gain is calculated as\n",
    "\n",
    "$$ \n",
    " Gain(P, P_{G_1}, P_{G_2}) = Entropy(P) - \\sum_{v \\epsilon\\{G_1,G_2 \\}} \\frac{|P_v|}{|P|}Entropy(P_v)\n",
    "$$\n",
    "\n",
    ">where Entropy is calculated as\n",
    "$$\\sum_{i=1}^c -p_i log_2 p_i$$\n",
    "\n",
    "> The peaks are then sorted based on the informatin gain and the user can choose the cutoff value for selecting the DA peaks. A suggested value for this would be values < 0.01\n",
    "\n",
    "\n",
    "`Usage`:\n",
    "> `getDiffAccessLogsitic(count_data, groups)`\n",
    "\n",
    "Arguements:\n",
    "> `cdBinary`: The peak information converted into binary. Ideally this would be file that set in the `outputPeakFileName`. `cdBinary` would be used to calculate the Jaccard distance which would then be used as d matrix in MDS calculation<br>\n",
    "\n",
    ">`groups` : groups into which each cell belong to.\n",
    "\n",
    "Value:\n",
    "> Returns a table with the peaks along with Information gain."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "%%R\n",
    "getDiffAccess <- function(cdBinary, groups=NULL){\n",
    "\n",
    "    if (length(levels(groups)) != 2) {\n",
    "        stop(paste(\"ERROR: wrong number of levels in the grouping factor (\", \n",
    "            paste(levels(groups), collapse = \" \"), \"), but must be two.\", \n",
    "            sep = \"\"))\n",
    "    }\n",
    "    if (is.null(groups)){\n",
    "        stop(\"ERROR: groups factor is not provided\")\n",
    "    }\n",
    "    \n",
    "    \n",
    "    SingleCell.Group1.CellNames <- names(groups[groups==levels(groups)[1]])\n",
    "    SingleCell.Group2.CellNames <- names(groups[groups==levels(groups)[2]])\n",
    "    \n",
    "    SingleCell.Group1.Binary <- cdBinary[,SingleCell.Group1.CellNames]\n",
    "    SingleCell.Group2.Binary <- cdBinary[,SingleCell.Group2.CellNames]\n",
    "\n",
    "    SingleCell.Group1.ZeroCount <- rowSums(SingleCell.Group1.Binary==0)\n",
    "    SingleCell.Group2.ZeroCount <- rowSums(SingleCell.Group2.Binary==0)\n",
    "    SingleCell.Group1.OneCount <- rowSums(SingleCell.Group1.Binary==1)\n",
    "    SingleCell.Group2.OneCount <- rowSums(SingleCell.Group2.Binary==1)\n",
    "    \n",
    "    SingleCell.Group1VsGroup2 <- data.frame(Chr = cdBinary[,1],\n",
    "                                           Start = cdBinary[,2],\n",
    "                                           end = cdBinary[,3])\n",
    "\n",
    "    SingleCell.Group1VsGroup2$group1OneCounts <- SingleCell.Group1.OneCount\n",
    "    SingleCell.Group1VsGroup2$group2OneCounts <- SingleCell.Group2.OneCount\n",
    "    SingleCell.Group1VsGroup2$log2Fold <- log2((SingleCell.Group1.OneCount+1)/dim(SingleCell.Group1.Binary)[2]) - \n",
    "                                            log2(((SingleCell.Group2.OneCount+1)/dim(SingleCell.Group2.Binary)[2]))\n",
    "    pvalue = vector(mode=\"numeric\", length=length(SingleCell.Group2.OneCount))\n",
    "    nGenes <- nrow(SingleCell.Group1.Binary)\n",
    "    \n",
    "    \n",
    "    for(i in 1:length(SingleCell.Group2.OneCount))\n",
    "    {\n",
    "        contingencyTable <- matrix(c(SingleCell.Group1.ZeroCount[i], SingleCell.Group2.ZeroCount[i],\n",
    "                                     SingleCell.Group1.OneCount[i], SingleCell.Group2.OneCount[i]), ncol=2)\n",
    "        SingleCell.Group1VsGroup2$pvalue[i] <- fisher.test(contingencyTable)$p.value        \n",
    "    }\n",
    "    \n",
    "    # Bonferroni correction\n",
    "    SingleCell.Group1VsGroup2$p_adjust <- p.adjust(SingleCell.Group1VsGroup2$pvalue, method=\"bonferroni\")\n",
    "\n",
    "    write.csv(SingleCell.Group1VsGroup2[order(SingleCell.Group1VsGroup2$p_adjust, decreasing=FALSE),], \n",
    "              paste0(levels(groups)[1],'_vs_',levels(groups)[2],'.csv'), row.names=FALSE)\n",
    "    \n",
    "    return(SingleCell.Group1VsGroup2)    \n",
    "}"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "`getDiffAccess()`:\n",
    "> Caculates differential accessibility between two groups of cells. For each peak it creates a contingency table with number of 0's and 1's for the two groups. <br>\n",
    "\n",
    ">|               \t| Group-1 \t| Group-2 \t|\n",
    "|---------------\t|---------\t|---------\t|\n",
    "| Number of 0's \t|         \t|         \t|\n",
    "| Number of 1's \t|         \t|         \t| \n",
    "<br>\n",
    "It then runs Fisher exact test to calculate the p-values. These p-values are then corrected for multiple testing with Bonferroni correction. Finally the function returns a table with the peaks, their count of 1's in each group, the p-value and the q-value (corrected p-value).\n",
    "\n",
    "Arguements:\n",
    "> `cdBinary`: The peak information converted into binary. Ideally this would be file that set in the `outputPeakFileName`. `cdBinary` would be used to calculate the Jaccard distance which would then be used as d matrix in MDS calculation<br>\n",
    "\n",
    "Value:\n",
    "> Returns a table with the peaks, their count of 1's in each group, the p-value and the q-value (corrected p-value)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "%%R\n",
    "getPeakVsPrecision <- function(peakA, peakB,start=100,stepSize=100){\n",
    "    \n",
    "    if (colnames(peakA)[1] != 'chr' | colnames(peakA)[2] != 'start' | colnames(peakA)[3] != 'end') {\n",
    "        stop(paste(\"ERROR: Column name should have 'chr' 'start' 'end'\"))\n",
    "    }\n",
    "\n",
    "    dimToRun <- min(dim(peakA)[1],dim(peakB)[1])\n",
    "    if(dim(peakB)[1]>dimToRun){\n",
    "        peakB <- peakB[1:dimToRun,]\n",
    "    }\n",
    "    GenomicRangeObjA <- makeGRangesFromDataFrame(peakA,\n",
    "                         keep.extra.columns=FALSE,\n",
    "                         ignore.strand=TRUE,\n",
    "                         seqinfo=NULL,\n",
    "                         seqnames.field=c(\"chr\"),\n",
    "                         start.field=\"start\",\n",
    "                         end.field=c(\"end\"),\n",
    "                         starts.in.df.are.0based=FALSE)\n",
    "    j=1\n",
    "    Recall <- vector(mode=\"numeric\", length=0)\n",
    "    Precision <- vector(mode=\"numeric\", length=0)\n",
    "    for(i in seq(100,(dimToRun-stepSize),stepSize)){\n",
    "           GenomicRangeObjB <- makeGRangesFromDataFrame(peakB[1:i,],\n",
    "                         keep.extra.columns=FALSE,\n",
    "                         ignore.strand=TRUE,\n",
    "                         seqinfo=NULL,\n",
    "                         seqnames.field=c(\"chr\"),\n",
    "                         start.field=\"start\",\n",
    "                         end.field=c(\"end\"),\n",
    "                         starts.in.df.are.0based=FALSE)\n",
    "\n",
    "           write.table(GenomicRangeObjA, 'tmpPeakA.bed', sep='\\t', row.names=FALSE, quote=FALSE, col.names=FALSE)\n",
    "\t   write.table(GenomicRangeObjB, 'tmpPeakB.bed', sep='\\t', row.names=FALSE, quote=FALSE, col.names=FALSE)\n",
    "           bedToolsCmd = 'bedtools intersect -a tmpPeakB.bed -b tmpPeakA.bed -u'\n",
    "           intersectRes = system(bedToolsCmd, intern=TRUE)\n",
    "           \n",
    "           FP = dim(peakB[1:i,])[1] - length(intersectRes)\n",
    "           TP = length(intersectRes)\n",
    "           \n",
    "           GenomicRangeObjB <- makeGRangesFromDataFrame(peakB[(i+1):dim(peakB)[1],],\n",
    "                         keep.extra.columns=FALSE,\n",
    "                         ignore.strand=TRUE,\n",
    "                         seqinfo=NULL,\n",
    "                         seqnames.field=c(\"Chr\"),\n",
    "                         start.field=\"Start\",\n",
    "                         end.field=c(\"End\"),\n",
    "                         starts.in.df.are.0based=FALSE)\n",
    "                         \n",
    "           write.table(GenomicRangeObjA, 'tmpPeakA.bed', sep='\\t', row.names=FALSE, quote=FALSE, col.names=FALSE)\n",
    "\t   write.table(GenomicRangeObjB, 'tmpPeakB.bed', sep='\\t', row.names=FALSE, quote=FALSE, col.names=FALSE)\n",
    "           bedToolsCmd = 'bedtools intersect -a tmpPeakB.bed -b tmpPeakA.bed -u'\n",
    "           intersectRes = system(bedToolsCmd, intern=TRUE)\n",
    "\n",
    "           TN = dim(peakB[(i+1):dim(peakB)[1],]) - length(intersectRes)\n",
    "           FN = length(intersectRes)\n",
    "           Recall[j] = 100*(TP/(TP + FN))\n",
    "           Precision[j] = 100*(TP / (TP + FP))\n",
    "           j = j +1\n",
    "           \n",
    "   \n",
    "    }\n",
    "    df<-data.frame(x=Recall,y=Precision)\n",
    "    p <- ggplot(df, aes_string(x=\"x\",y =\"y\"))+ scale_colour_hue(l=40)\n",
    "    \n",
    "    p <- p + ggtitle(\"Recall vs Precision\") + theme(plot.title = element_text(size = 16, face = \"bold\"))\n",
    "    p <- p + geom_point(size = 2)\n",
    "    p <- p + xlab(\"Recall%\") \n",
    "    p <- p + ylab(\"Precision%\")+\n",
    "\ttheme_light(base_size=15) +\n",
    "        theme(strip.background = element_blank(),\n",
    "              panel.border     = element_blank(),\n",
    "              plot.title = element_text(hjust = 0.5))\n",
    "    print(p)\n",
    "    \n",
    "    #GenomicRangeObjB <- makeGRangesFromDataFrame(peakB,\n",
    "    #                     keep.extra.columns=FALSE,\n",
    "    #                     ignore.strand=TRUE,\n",
    "    #                     seqinfo=NULL,\n",
    "    #                     seqnames.field=c(\"Chr\"),\n",
    "    #                     start.field=\"Start\",\n",
    "    #                     end.field=c(\"End\"),\n",
    "    #                     starts.in.df.are.0based=FALSE)\n",
    "    #                     \n",
    "    # write.table(GenomicRangeObjA, 'tmpPeakA.bed', sep='\\t', row.names=FALSE, quote=FALSE, col.names=FALSE)\n",
    "    # write.table(GenomicRangeObjB, 'tmpPeakB.bed', sep='\\t', row.names=FALSE, quote=FALSE, col.names=FALSE)\n",
    "    # bedToolsCmd = 'bedtools intersect -a tmpPeakB.bed -b tmpPeakA.bed -u'\n",
    "    # intersectRes = system(bedToolsCmd, intern=TRUE)\n",
    "   \n",
    "   #return(intersectRes) \n",
    "}"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "`getPeakVsPrecision()`:\n",
    "> This function is used to see how many peaks from the Bulk data the single-cell have been able to capture. For that we calculate the Peak and the Precision of the single-cell data against its Bulk counterpart and then plot it. The way these two values are calculated are\n",
    "\n",
    "$$\n",
    "Recall = 100  \\frac{TP}{TP + FN}\n",
    "$$\n",
    "$$\n",
    "Precision = 100  \\frac{TP}{TP + FP}\n",
    "$$\n",
    "> where <br>\n",
    "> TP = True Positive <br>\n",
    "> TN = True Negative <br>\n",
    "> FP = False Positive <br>\n",
    "> FN = False Negative <br>\n",
    "\n",
    "`Usage`:\n",
    "> `getPeakVsPrecision(peakA, peakB,start,stepSize)`\n",
    "\n",
    "Arguements:\n",
    "> `peakA`: Peak of the BULK data <br>\n",
    "> `peakB`: Single-cell Peak <br>\n",
    "> `start`: Number of peaks to start with the calculation, default is `100`. <br>\n",
    "> `stepSize`: Size of the increase steps\n",
    "\n",
    "Value:\n",
    "> Plots Recall vs. Precision and returns a summary"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "%%R\n",
    "getJaccardDist <- function(cdBinary){\n",
    "        \n",
    "    if(colnames(cdBinary[,2:3])[1] == 'start' && colnames(cdBinary[,2:3])[2] == 'end'){\n",
    "        SingleCell.Binary <- cdBinary[,4:(dim(cdBinary)[2])]\n",
    "    }\n",
    "    else\n",
    "        SingleCell.Binary <- cdBinary\n",
    "    \n",
    "    \n",
    "    SingleCell.Binary.Jaccard <- jaccard(as.matrix(SingleCell.Binary))\n",
    "    \n",
    "    return(SingleCell.Binary.Jaccard)\n",
    "}"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "`getJaccardDist()`:\n",
    "> This function calculates the jaccard distance between each of the cells. Jaccard distance calculates the ratio between the differences of the sizes of the union and intersection against the size of the union. The equation for calculating this distance is <br><br>$$d_J(A,B)=\\frac{|A \\cup B| - |A \\cap B|}{|A \\cup B|}$$\n",
    "\n",
    "Arguements:\n",
    "> `cdBinary`: The peak information converted into binary. Ideally this would be file that set in the `outputPeakFileName`. \n",
    "\n",
    "\n",
    "Value:\n",
    "> Jaccard distance between each of the cell."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [],
   "source": [
    "%%R\n",
    "plotMDS <- function(cdBinary, k, groups=NULL, cellName, ret.val=FALSE, text.label=FALSE, \n",
    "                    title=\"\", coordinates=c(1,2)){\n",
    "    if(missing(ret.val)){\n",
    "        ret.val = FALSE\n",
    "    }\n",
    "    if(text.label==TRUE & missing(cellName)){\n",
    "        stop(\"ERROR: Please give cellName\")\n",
    "    }\n",
    "    \n",
    "    SingleCell.Binary.Jaccard <- getJaccardDist(cdBinary)\n",
    "    fit <- cmdscale(as.dist(SingleCell.Binary.Jaccard),eig=TRUE, k=k)\n",
    "\n",
    " \n",
    "    if(is.null(groups)){\n",
    "        df<-data.frame(x=fit$points[,coordinates[1]],y=fit$points[,coordinates[2]], Cell=colnames(SingleCell.Binary.Jaccard))\n",
    "        p <- ggplot(df, aes_string(x=\"x\",y =\"y\"))+ scale_colour_hue(l=40)\n",
    "    }\n",
    "    else{\n",
    "        df<-data.frame(x=fit$points[,coordinates[1]],y=fit$points[,coordinates[2]], Cell=colnames(SingleCell.Binary.Jaccard), Batch=groups)\n",
    "         p <- ggplot(df, aes_string(x=\"x\",y =\"y\", color=\"Batch\"))+ \n",
    "            scale_colour_hue(l=40) + \n",
    "            scale_colour_brewer(palette=\"Set1\") \n",
    "    }\n",
    "\n",
    "    \n",
    "    p <- p + ggtitle(title) + theme(plot.title = element_text(size = 20, face = \"bold\"))\n",
    "    p <- p + geom_point(size = 2)\n",
    "    p <- p + xlab(paste0(\"Coordinate \",coordinates[1])) \n",
    "    p <- p + ylab(paste0(\"Coordinate \",coordinates[2]))+\n",
    "\ttheme_light(base_size=24) +\n",
    "        theme(strip.background = element_blank(),\n",
    "              panel.border     = element_blank(),\n",
    "              plot.title = element_text(hjust = 0.5))\n",
    "\n",
    "       \n",
    "    if(text.label==TRUE)\n",
    "        p <- p + geom_text(data=df,aes(label=cellName),alpha=0.5,size=4, vjust=1,hjust=0.5,angle=45)\n",
    "\n",
    "    #p <- p + theme(axis.title = element_text(size = 14), axis.text = element_text(size = 14),\n",
    "    #legend.text = element_text(size = 14), legend.title = element_text(size = 14))\n",
    "    print(p)\n",
    "    \n",
    "    if(ret.val == TRUE)\n",
    "        return(fit)\n",
    "}\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "`plotMDS()`:\n",
    "> This function calculates the Multidimensional Scaling (MDS) and plots the cells based on Jaccard distance to visualize the level of similarity between the cells.\n",
    "\n",
    "Arguements:\n",
    "> `cdBinary`: The peak information converted into binary. Ideally this would be file that set in the `outputPeakFileName`. `cdBinary` would be used to calculate the Jaccard distance which would then be used as d matrix in MDS calculation<br>\n",
    "`k`: The maximum dimension of the space which the data are to be represented in; must be in {1, 2, ..., n-1}. <br>\n",
    "`groups`: If the cells can be organized into different groups (for eg. different batches), then they are passed as groups making them factor variable.<br>\n",
    "`cellName`: If `text.label` is set to `TRUE` then you have to supply `cellName`.\n",
    "`ret.val`: Need to set this TRUE for the function to return the MDS values <br>\n",
    "`text.label`: Need to set this TRUE if label of each of the cells are to be printed in the plot.\n",
    "`coordinates`: The coordinates to be plotted for the MDS plot\n",
    "\n",
    "\n",
    "Value:\n",
    "> Calculates the MDS and plots the cells based on their MDS dimension. It returns the MDS values if `ret.val` is set to TRUE."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [],
   "source": [
    "%%R\n",
    "plotMDSClust <- function(cdBinary, k, Clusters=NULL, cellName, ret.val=FALSE, text.label=FALSE, \n",
    "                    title=\"\"){\n",
    "\n",
    "    if(missing(k)){\n",
    "        stop(\"ERROR: Number of Clusters \\\"k\\\" is missing\")\n",
    "    }\n",
    "    if(missing(ret.val)){\n",
    "        ret.val = FALSE\n",
    "    }\n",
    "    if(text.label==TRUE & missing(cellName)){\n",
    "        stop(\"ERROR: Please give cellName\")\n",
    "    }\n",
    "    \n",
    "    SingleCell.Binary.Jaccard <- getJaccardDist(cdBinary)\n",
    "    fit <- cmdscale(as.dist(SingleCell.Binary.Jaccard),eig=TRUE, k=k)\n",
    " \n",
    "    if(is.null(Clusters)){\n",
    "        df<-data.frame(x=fit$points[,1],y=fit$points[,2], Cell=colnames(SingleCell.Binary.Jaccard))\n",
    "        p <- ggplot(df, aes_string(x=\"x\",y =\"y\"))+ scale_colour_hue(l=40)\n",
    "    }\n",
    "    else{\n",
    "        df<-data.frame(x=fit$points[,1],y=fit$points[,2], Cell=colnames(SingleCell.Binary.Jaccard), Clusters=Clusters)\n",
    "        p <- ggplot(df, aes_string(x=\"x\",y =\"y\", color=\"Clusters\"))+ scale_colour_hue(l=40)\n",
    "    }\n",
    "\n",
    "    \n",
    "    p <- p + ggtitle(title) + theme(plot.title = element_text(size = 20, face = \"bold\"))\n",
    "    p <- p + geom_point(size = 2)\n",
    "    p <- p + xlab(\"Coordinate 1\") \n",
    "    p <- p + ylab(\"Coordinate 2\")+\n",
    "\ttheme_light(base_size=24) +\n",
    "        theme(strip.background = element_blank(),\n",
    "              panel.border     = element_blank(),\n",
    "              plot.title = element_text(hjust = 0.5))\n",
    "\n",
    "    if(length(levels(Clusters)) < 12)\n",
    "        p <- p + scale_colour_brewer(palette=\"Set1\")        \n",
    "    if(text.label==TRUE)\n",
    "        p <- p + geom_text(data=df,aes(label=cellName),alpha=0.5,size=4, vjust=1,hjust=0.5,angle=45)\n",
    "\n",
    "    #p <- p + theme(axis.title = element_text(size = 14), axis.text = element_text(size = 14),\n",
    "    #legend.text = element_text(size = 14), legend.title = element_text(size = 14))\n",
    "    \n",
    "    print(p)\n",
    "    \n",
    "    if(ret.val == TRUE)\n",
    "        return(fit)\n",
    "}"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "`plotMDSClust()`:\n",
    "> This function calculates the Multidimensional Scaling (MDS) and plots the cells based on Jaccard distance to visualize the level of similarity between the cells. It is very similar to `plotMDS()` except we colour it based on clusters.\n",
    "\n",
    "Arguements:\n",
    "> `cdBinary`: The peak information converted into binary. Ideally this would be file that set in the `outputPeakFileName`. `cdBinary` would be used to calculate the Jaccard distance which would then be used as d matrix in MDS calculation<br>\n",
    "`k`: The maximum dimension of the space which the data are to be represented in; must be in {1, 2, ..., n-1}. <br>\n",
    "`groups`: If the cells can be organized into different groups (for eg. different batches), then they are passed as groups making them factor variable.<br>\n",
    "`cellName`: If `text.label` is set to `TRUE` then you have to supply `cellName`.\n",
    "`ret.val`: Need to set this TRUE for the function to return the MDS values <br>\n",
    "`text.label`: Need to set this TRUE if label of each of the cells are to be printed in the plot.\n",
    "\n",
    "\n",
    "Value:\n",
    "> Calculates the MDS and plots the cells based on their MDS dimension. It returns the MDS values if `ret.val` is set to TRUE."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "%%R\n",
    "plotPCAJaccard <- function(cdBinary, groups=NULL, ret.val=FALSE , cellName, text.label=FALSE, title){\n",
    "    \n",
    "   \n",
    "    if(missing(ret.val)){\n",
    "        ret.val = FALSE\n",
    "    }\n",
    "    if(text.label==TRUE & missing(cellName)){\n",
    "        stop(\"ERROR: Please give cellName\")\n",
    "    }\n",
    "    \n",
    "    SingleCell.Binary.Jaccard <- getJaccardDist(cdBinary)\n",
    "    FinalPCAData <- t(SingleCell.Binary.Jaccard)\n",
    "    PCx=1\n",
    "    PCy=2\n",
    "    pcaPRComp <- prcomp(FinalPCAData)\n",
    "    percentVar <- pcaPRComp$sdev^2/sum(pcaPRComp$sdev^2)\n",
    " \n",
    "    if(is.null(groups)){\n",
    "        df<-data.frame(PCX=pcaPRComp$x[,PCx],PCY=pcaPRComp$x[,PCy])\n",
    "        p1<-ggplot(df, aes_string(x=\"PCX\",y =\"PCY\"))        \n",
    "    }\n",
    "    else{\n",
    "        df<-data.frame(PCX=pcaPRComp$x[,PCx],PCY=pcaPRComp$x[,PCy], groups=groups)\n",
    "        p1<-ggplot(df, aes_string(x=\"PCX\",y =\"PCY\", color=\"groups\"))             \n",
    "    }          \n",
    "    if(missing(title))\n",
    "        p1<-p1+ggtitle(\"PCA with Jacard Matrix\")\n",
    "    else\n",
    "        p1<-p1+ggtitle(title)\n",
    "        \n",
    "    p1<-p1+geom_point(size = 3)\n",
    "    p1<-p1+xlab(paste(\"PC\",PCx,\": \", round(percentVar[PCx] * 100), \"% variance\"))\n",
    "    p1<-p1+ylab(paste(\"PC\",PCy,\": \", round(percentVar[PCy] * 100), \"% variance\"))\n",
    "    if(text.label == TRUE){\n",
    "        p1<-p1+geom_text(data=df,aes(label=cellName),\n",
    "                       alpha=0.5,size=4, vjust=1,hjust=0.5,angle=45, color=\"black\")\n",
    "    }\n",
    "    \n",
    "    p1<-p1 +\n",
    "    theme_light(base_size=15)+     \n",
    "         theme(plot.title = element_text(hjust = 0.5),\n",
    "              panel.border     = element_blank())\n",
    "\n",
    "    if(length(levels(groups)) < 12)\n",
    "        p1 <- p1 + scale_colour_brewer(palette=\"Set1\")\n",
    "    \n",
    "    print(p1)\n",
    "        \n",
    "    if(ret.val == TRUE)\n",
    "        return(pcaPRComp)\n",
    "}"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "`plotPCAJaccard()`:\n",
    "> This function calculates the Principle Component Analysis (PCA) and plots the cells based on Jaccard distance in a reduced dimension to visualize how the cells cluster together.\n",
    "\n",
    "Arguements:\n",
    "> `cdBinary`: The peak information converted into binary. Ideally this would be file that set in the `outputPeakFileName`. `cdBinary` would be used to calculate the Jaccard distance which would then be used as d matrix in MDS calculation<br>\n",
    "`groups`: If the cells can be organized into different groups (for eg. different batches), then they are passed as groups making them factor variable.<br>\n",
    "`ret.val`: Need to set this TRUE for the function to return the MDS values <br>\n",
    "`text.label`: Need to set this TRUE if label of each of the cells are to be printed in the plot.<br>\n",
    "`title`: Sets the title of the plot.\n",
    "\n",
    "\n",
    "Value:\n",
    "> Calculates PCA and plots the cells in reduced dimension. If`ret.val` is set to TRUE it returns the PCA components along with standard deviations. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "%%R\n",
    "plotMultiplePCAJaccard <- function(cdBinary, nPCAToDisplay, groups=NULL, ret.val=FALSE , text.label=FALSE, title=\"\"){\n",
    "    \n",
    "    if(missing(nPCAToDisplay)){\n",
    "    \tstop(\"ERROR: Number of PCA's \\\"nPCAToDisplay\\\" to display is missing\")\n",
    "    }\n",
    "    if(missing(ret.val)){\n",
    "        ret.val = FALSE\n",
    "    }\n",
    "    \n",
    "    SingleCell.Binary.Jaccard <- getJaccardDist(cdBinary)\n",
    "    FinalPCAData <- t(SingleCell.Binary.Jaccard)\n",
    "    PCx=1\n",
    "    PCy=2\n",
    "    pcaPRComp <- prcomp(FinalPCAData)\n",
    "\n",
    "    nmax = nPCAToDisplay\n",
    "    plotTitle = paste0('Plotting first ',nmax,' PCAs')\n",
    "    \n",
    "    txt1 <- paste(\"Percent_PC_Var_onfirst\",nmax,\"PCs\",sep=\"\")\n",
    "    pca_var = pcaPRComp$sdev ^ 2\n",
    "    pca_var_percent <- 100 * pca_var / sum(pca_var)\n",
    "    pca_var_percent_first10 <- NA * pca_var\n",
    "    pca_var_percent_first10[1:nmax] <- 100 * pca_var[1:nmax] / sum(pca_var[1:nmax])\n",
    "\n",
    "    pca_var_out <- data.frame(round(pca_var,3),round(pca_var_percent,1),\n",
    "                          round(pca_var_percent_first10,1))\n",
    "    rownames(pca_var_out) <- colnames(pcaPRComp$x)\n",
    "    colnames(pca_var_out) <- c(\"PC_Var\",\"PC_Var_percent\",txt1)\n",
    "\n",
    "\n",
    "    if(is.null(groups)){\n",
    "        df <- as.data.frame(pcaPRComp$x)\n",
    "        p <- ggpairs(df, columns=1:nPCAToDisplay, upper=list(continuous=\"points\"), \n",
    "             title=plotTitle,              \n",
    "             columnLabels = as.character(paste0(colnames(df[,1:nPCAToDisplay]), ' : ', \n",
    "                                                pca_var_out$PC_Var_percent[1:nPCAToDisplay], '% variance')))+\n",
    "    \t     theme_light(base_size=15)+     \n",
    "             theme(plot.title = element_text(hjust = 0.5))        \n",
    "    }\n",
    "    else{\n",
    "        df <- as.data.frame(pcaPRComp$x)\n",
    "        df$Cell=groups        \n",
    "        p <- ggpairs(df, columns=1:nPCAToDisplay, upper=list(continuous=\"points\"), \n",
    "             title=plotTitle, \n",
    "             mapping = aes_string(color=\"Cell\"),\n",
    "             legend = c(1,nPCAToDisplay),\n",
    "             columnLabels = as.character(paste0(colnames(df[,1:nPCAToDisplay]), ' : ', \n",
    "                                                pca_var_out$PC_Var_percent[1:nPCAToDisplay], '% variance')))+\n",
    "             theme_light(base_size=15)+     \n",
    "             theme(plot.title = element_text(hjust = 0.5))\n",
    "    }          \n",
    "    \n",
    "    print(p)\n",
    "}"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "`plotMultiplePCAJaccard()`:\n",
    "> This function calculates the Principle Component Analysis (PCA) based on Jacard distance and plots the first `n` components of the PCA.\n",
    "\n",
    "Arguements:\n",
    "> `cdBinary`: The peak information converted into binary. Ideally this would be file that set in the `outputPeakFileName`. `cdBinary` would be used to calculate the Jaccard distance which would then be used as d matrix in MDS calculation<br>\n",
    "`nPCAToDisplay` : Number of PCAs to display. <br>\n",
    "`groups`: If the cells can be organized into different groups (for eg. different batches), then they are passed as groups making them factor variable.<br>\n",
    "`ret.val`: Need to set this TRUE for the function to return the MDS values <br>\n",
    "`text.label`: Need to set this TRUE if label of each of the cells are to be printed in the plot.<br>\n",
    "`title`: Sets the title of the plot.\n",
    "\n",
    "\n",
    "Value:\n",
    "> Calculates PCA and plots the cells in first `n` components. If`ret.val` is set to TRUE it returns the PCA components along with standard deviations. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "%%R\n",
    "plotVarExplained <- function(cdBinary, nPCAToDisplay, groups=NULL, ret.val=FALSE , text.label=FALSE, title=\"\"){\n",
    "\n",
    "    if(missing(nPCAToDisplay)){\n",
    "    \tstop(\"ERROR: Number of PCA's \\\"nPCAToDisplay\\\" to display is missing\")\n",
    "    }\n",
    "    if(missing(ret.val)){\n",
    "        ret.val = FALSE\n",
    "    }\n",
    "    \n",
    "    SingleCell.Binary.Jaccard <- getJaccardDist(cdBinary)\n",
    "    FinalPCAData <- t(SingleCell.Binary.Jaccard)\n",
    "    PCx=1\n",
    "    PCy=2\n",
    "    pcaPRComp <- prcomp(FinalPCAData)\n",
    "\n",
    "    nmax = nPCAToDisplay\n",
    "    plotTitle = paste0('Plotting first ',nmax,' PCAs')\n",
    "    \n",
    "    txt1 <- paste(\"Percent_PC_Var_onfirst\",nmax,\"PCs\",sep=\"\")\n",
    "    pca_var = pcaPRComp$sdev ^ 2\n",
    "    pca_var_percent <- 100 * pca_var / sum(pca_var)\n",
    "    pca_var_percent_first10 <- NA * pca_var\n",
    "    pca_var_percent_first10[1:nmax] <- 100 * pca_var[1:nmax] / sum(pca_var[1:nmax])\n",
    "\n",
    "    pca_var_out <- data.frame(round(pca_var,3),round(pca_var_percent,1),round(pca_var_percent_first10,1))\n",
    "    rownames(pca_var_out) <- colnames(pcaPRComp$x)\n",
    "    colnames(pca_var_out) <- c(\"PC_Var\",\"PC_Var_percent\",\"PC_Var_n_percent\")\n",
    "\n",
    "\n",
    "    p <- ggplot(pca_var_out[1:nmax,]) + scale_colour_hue(l=40) \n",
    "    p <- p + geom_line(aes(x=c(1:nmax), y = PC_Var_percent, color=\"PC_Var_percent\"), size=1.5)+\n",
    "         geom_point(aes(x=c(1:nmax), y = PC_Var_percent, color=\"PC_Var_percent\"), size = 2)+\n",
    "         geom_line(aes(x=c(1:nmax), y = PC_Var_n_percent, color=\"PC_Var_n_percent\"), size=1.5)+\n",
    "         geom_point(aes(x=c(1:nmax), y = PC_Var_n_percent, color=\"PC_Var_n_percent\"), size = 2)+\n",
    "         ggtitle(\"Total variance by the PCAs considered\")+\n",
    "         xlab(paste(\"PC1 to PC10\"))+\n",
    "         ylab(paste(\"% of Variance\"))+     \n",
    "         theme_light(base_size=15)+     \n",
    "         theme(plot.title = element_text(hjust = 0.5),\n",
    "              panel.border     = element_blank())+\n",
    "         scale_x_continuous(breaks=c(1:nmax))+\n",
    "         scale_colour_brewer(palette=\"Set1\")\n",
    "         \n",
    "   print(p)\n",
    "}"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "`plotVarExplained()`:\n",
    "> This function calculates the percent of variance explained by the PCAs and plots it.\n",
    "\n",
    "Arguements:\n",
    "> `cdBinary`: The peak information converted into binary. Ideally this would be file that set in the `outputPeakFileName`. `cdBinary` would be used to calculate the Jaccard distance which would then be used as d matrix in MDS calculation<br>\n",
    "`nPCAToDisplay` : Number of PCAs to display. <br>\n",
    "`groups`: If the cells can be organized into different groups (for eg. different batches), then they are passed as groups making them factor variable.<br>\n",
    "`ret.val`: Need to set this TRUE for the function to return the MDS values <br>\n",
    "`text.label`: Need to set this TRUE if label of each of the cells are to be printed in the plot.<br>\n",
    "`title`: Sets the title of the plot.\n",
    "\n",
    "\n",
    "Value:\n",
    "> Calculates the percent of variance explained by the PCAs and the plots thsoe percentage. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "%%R\n",
    "plot_tSNE <- function(cdBinary, nDimToUSE, groups=NULL, cellName, perplexity_division, ret.val=FALSE , text.label=FALSE, title=\"\"){\n",
    "\n",
    "    if(missing(nDimToUSE)){\n",
    "    \tstop(\"ERROR: Number of PCA's \\\"nDimToUSE\\\" to use is missing\")\n",
    "    }\n",
    "    if(missing(ret.val)){\n",
    "        ret.val = FALSE\n",
    "    }\n",
    "    \n",
    "    if(text.label==TRUE & missing(cellName)){\n",
    "        stop(\"ERROR: Please give cellName\")\n",
    "    }\n",
    "    \n",
    "    if(missing(perplexity_division)){\n",
    "        stop(\"ERROR: Please enter the number with which you want to divide the dimension of your data for perplexity setting\")\n",
    "    }\n",
    "        \n",
    "    \n",
    "    #SingleCell.Binary.Jaccard <- getJaccardDist(cdBinary)\n",
    "    #FinalPCAData <- t(SingleCell.Binary.Jaccard)\n",
    "    \n",
    "    SingleCell.Binary.Jaccard <- getJaccardDist(cdBinary)\n",
    "    fit <- cmdscale(as.dist(SingleCell.Binary.Jaccard),eig=TRUE, k=nDimToUSE)\n",
    "\n",
    "    #pcaPRComp <- prcomp(FinalPCAData)\n",
    "\n",
    "    \n",
    "    #rtsne_pca_out <- Rtsne(as.matrix(pcaPRComp$x[,1:nPCAToUSE]), perplexity = dim(pcaPRComp$x)[1]/perplexity_division, check_duplicates = FALSE)\n",
    "    rtsne_pca_out <- Rtsne(as.matrix(fit$points[,1:nDimToUSE]), perplexity = perplexity_division, \n",
    "                           check_duplicates = FALSE, pca=FALSE, theta=0.01, max_iter=3000)\n",
    "    \n",
    "    if(is.null(groups)){\n",
    "        if(ret.val==TRUE)\n",
    "            df<-data.frame(X=rtsne_pca_out$Y[,1],Y=rtsne_pca_out$Y[,2], cellName=cellName)\n",
    "        else\n",
    "            df<-data.frame(X=rtsne_pca_out$Y[,1],Y=rtsne_pca_out$Y[,2])\n",
    "        p1 <- ggplot(df, aes_string(x=\"X\",y =\"Y\"))\n",
    "    }\n",
    "    else{\n",
    "        if(ret.val==TRUE)\n",
    "            df<-data.frame(X=rtsne_pca_out$Y[,1],Y=rtsne_pca_out$Y[,2], Cell=colnames(SingleCell.Binary.Jaccard), Batch=groups, cellName=cellName)\n",
    "        else\n",
    "            df<-data.frame(X=rtsne_pca_out$Y[,1],Y=rtsne_pca_out$Y[,2], Cell=colnames(SingleCell.Binary.Jaccard), Batch=groups)\n",
    "        p1 <- ggplot(df, aes_string(x=\"X\",y =\"Y\", color=\"Batch\"))\n",
    "    }\n",
    "    \n",
    "\n",
    "    p1<-p1 + ggtitle(\"t-SNE plot\")\n",
    "    p1<-p1 + geom_point(size = 2) \n",
    "    p1<-p1 + xlab(paste(\"Dim-1\"))\n",
    "    p1<-p1 + ylab(\"Dim-2\")+\n",
    "        theme_light(base_size=24) +\n",
    "        theme(strip.background = element_blank(),\n",
    "              panel.border     = element_blank(),\n",
    "              plot.title = element_text(hjust = 0.5))\n",
    "    if(length(levels(groups)) < 12)\n",
    "        p1 <- p1 + scale_colour_brewer(palette=\"Set1\")\n",
    "    if(text.label==TRUE)\n",
    "        p1<-p1 + geom_text(data=df,aes(label=cellName),alpha=0.5,size=4, vjust=1,hjust=0.5,angle=45)\n",
    "    print(p1)\n",
    "    return(df)\n",
    "}"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "`plot_tSNE()`:\n",
    "> Caculates the t-distributed stochastic neighbor embedding (t-SNE) which is a non-linear dimensionality reduction technique particualrly suitable for embedding high-dimensional data into a space of two or three dimensions. This function then plots the lower dimensional representation of the cells in a 2-D scatter plot.\n",
    "\n",
    "Arguements:\n",
    "> `cdBinary`: The peak information converted into binary. Ideally this would be file that set in the `outputPeakFileName`. `cdBinary` would be used to calculate the Jaccard distance which would then be used as d matrix in MDS calculation<br>\n",
    "`nPCAToDisplay` : Number of PCAs to be used to calcualte the t-SNE values. <br>\n",
    "`groups`: If the cells can be organized into different groups (for eg. different batches), then they are passed as groups making them factor variable.<br>\n",
    "`ret.val`: Need to set this TRUE for the function to return the MDS values <br>\n",
    "`text.label`: Need to set this TRUE if label of each of the cells are to be printed in the plot.<br>\n",
    "`title`: Sets the title of the plot.\n",
    "\n",
    "\n",
    "Value:\n",
    "> Calculates and plots the t-SNE representation of the cells. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "%%R\n",
    "clean.count.peaks <- function(cd, min.cell.peaks.obs = 10, min.peaks.cell = 50){\n",
    "    cdFilt <- cd[,1:3]\n",
    "    cdCells <- cd[,4:dim(cd)[2]]\n",
    "    valid.cell.ix <- which(colSums(cdCells) >= min.peaks.cell)\n",
    "    cdFilt <- cbind(cdFilt,cdCells[,valid.cell.ix])\n",
    "    cdFilt <- cdFilt[rowSums(cdFilt[,4:dim(cdFilt)[2]]) >=min.cell.peaks.obs ,]\n",
    "    cdFiltTemp <- cdFilt[,1:3]\n",
    "    cdCells <- cdFilt[,4:dim(cdFilt)[2]]\n",
    "    valid.cell.ix <- which(colSums(cdCells) > 0)\n",
    "    cdFilt <- cbind(cdFiltTemp,cdCells[,valid.cell.ix])\n",
    "    return(cdFilt)\n",
    "}"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "`clean.count.peaks()`:\n",
    "> This function filter peaks and cells based on peak and cell requirements.\n",
    "\n",
    "Usage:\n",
    "> `clean.count.peaks(cd, min.cell.peaks.obs = 10, min.peaks.cell = 50)`\n",
    "\n",
    "Arguements:\n",
    "> `cd`: A binay matrix with the peak accessiblity of all the single-cells. The row would reprsent each of the peaks unique name and the columns would be each cell. Each entry in the matrix is 0/1 representing either the peak is close/open for that cell. <br>\n",
    "> `min.cell.peaks.obs`: Minimum number of cells that a peak has to be observed in order for that peak to be a valid peak. <br>\n",
    "> `min.peaks.cell` : Minimum number of peaks that a cell needs to have in order for that cell to be called a valid cell. <br>\n",
    "\n",
    "Value:\n",
    "> A matrix with valid accessible peaks."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "%%R\n",
    "multiplot <- function(..., plotList=NULL, File, cols=1, layout=NULL) {\n",
    "\n",
    "  # Adopted from Cookbook for R\n",
    "  library(grid)\n",
    "\n",
    "  plots <- c(list(...), plotList)\n",
    "\n",
    "  numPlots = length(plots)\n",
    "\n",
    "  if (is.null(layout)) {\n",
    "    layout <- matrix(seq(1, cols * ceiling(numPlots/cols)),\n",
    "                    ncol = cols, nrow = ceiling(numPlots/cols))\n",
    "  }\n",
    "\n",
    " if (numPlots==1) {\n",
    "    print(plots[[1]])\n",
    "\n",
    "  } else {\n",
    "    grid.newpage()\n",
    "    pushViewport(viewport(layout = grid.layout(nrow(layout), ncol(layout))))\n",
    "\n",
    "    # Make each plot, in the correct location\n",
    "    for (i in 1:numPlots) {\n",
    "      # Get the i,j matrix positions of the regions that contain this subplot\n",
    "      matchidx <- as.data.frame(which(layout == i, arr.ind = TRUE))\n",
    "\n",
    "      print(plots[[i]], vp = viewport(layout.pos.row = matchidx$row,\n",
    "                                      layout.pos.col = matchidx$col))\n",
    "    }\n",
    "  }\n",
    "}"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "--------------"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Analysis\n",
    "Below is the analysis for the deconvolution approach"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Configuration Cell\n",
    "First we configure all the parameters for the Scasat tool."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "# This cell holds the parameters for the tools\n",
    "\n",
    "# Set this parameter to the folder where all the Peak Files are. This has to be the folder\n",
    "# where each of the individual cells peak files reside\n",
    "peakFolder = '/home/mqbsxsm2/Rna-seq_Data-Analysis/syed_scATAC_ReDoing/All_Filtered_MACS2'\n",
    "\n",
    "\n",
    "# Set this parameter to the folder where the peaks for the merge files are. This can be the same folder\n",
    "# of all the individual cells peak folder or can be a single peak file that is generated from merging BAM files\n",
    "peakFolderToMerge = '/home/mqbsxsm2/Rna-seq_Data-Analysis/syed_scATAC_ReDoing/All_Filtered_BAM/'\n",
    "\n",
    "\n",
    "\n",
    "# Set up the pattern for the BAM file ending\n",
    "BAMFilePattern = '_nodup_sorted_cleaned.bam'\n",
    "\n",
    "# Set this parameter to the pattern that is common to all the peak file names that you wish to merge. \n",
    "# for eg, if you want to combine OE19_C01_peaks.narrowPeak and OE19_C02_peaks.narrowPeak then the\n",
    "# pattern would be '_peaks.narrowPeak'\n",
    "# Allowed patterns while calling with the pipeline provided here are\n",
    "# _summits_shifted.bed and '_peaks.narrowPeak'\n",
    "peakFilePattern = '_peaks.narrowPeak'\n",
    "\n",
    "# This would be automatically set to 'mergedPeaks.bed' file in peakFolder folder\n",
    "mergedPeakFile = peakFolderToMerge + 'mergedPeaks.bed'\n",
    "\n",
    "# Set this folder to where the PeakHits.csv File would be saved\n",
    "outputFolder = peakFolderToMerge\n",
    "\n",
    "# Set the outputPeakFileName name here\n",
    "outputPeakFileName = 'MixedPopulation_rerun.csv'\n",
    "\n",
    "# This will be automatically set to the 'PeakHits.csv' file in outputFolder folder\n",
    "peakFile = outputFolder + outputPeakFileName"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We now call `getAggregatedPeak()` to merge all the BAM files and generated the aggregated peak from the merged BAM file."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [],
   "source": [
    "getAggregatedPeak(peakFolderToMerge, BAMFilePattern)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "To calculates the accessiblity of each of the peaks for all the cells we call the following function."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "%%R -i mergedPeakFile,peakFolder,peakFilePattern,outputFolder,outputPeakFileName\n",
    "peakAccessibility(mergedPeakFile, peakFolder, peakFilePattern, outputFolder, outputPeakFileName)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We now read the file listing all the peaks for all the single cells. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "%%R -i peakFile\n",
    "cd <- read.csv(peakFile, sep=' ', header=TRUE)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Filtering accessible peaks\n",
    "Now we filter all the low quality cells and peaks. This is a very important step. Setting the threshold too low would cause low quality cells and peaks to go into downstream anlaysis whereas setting the threshold too stringent would cause high quality cells to be dropped. The choice of parameters for this filtering depends on the experimental design for eg. cell types and the research question that is being asked."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[1] 236580    176\n",
       "[1] 52546   176\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%%R\n",
    "cdFiltTop <- clean.count.peaks(cd)\n",
    "print(dim(cd))\n",
    "print(dim(cdFiltTop))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [],
   "source": [
    "%%R\n",
    "write.csv(cdFiltTop, 'cdFiltTop_MixedPopulation_ReRun.csv', sep='\\t')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We seperate the cell names from the two batches"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "  [1] D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D\n",
       " [38] D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D\n",
       " [75] D D D D D D D D D D E E E E E E E E E E E E E E E E E E E E E E E E E E E\n",
       "[112] E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E\n",
       "[149] E E E E E E E E E E E E E E E E E E E E E E E E E\n",
       "Levels: D E\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%%R\n",
    "cellNames <- colnames(cdFiltTop[,4:dim(cdFiltTop)[2]])\n",
    "cellTypes <- as.factor(gsub('[0-9]','',cellNames))\n",
    "cellTypes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "cellTypes\n",
       " D  E \n",
       "84 89 \n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%%R\n",
    "table(cellTypes)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We now apply MDS and plot it in a 2-D plot."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n"
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%%R\n",
    "MDSClust <- plotMDS(cdFiltTop, k=2, groups=cellTypes, ret.val=TRUE, text.label=FALSE, title=\"\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "png \n",
       "  2 \n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%%R\n",
    "pdf('MDS_All_updated_Without_Filtering.pdf')\n",
    "MDSClust <- plotMDS(cdFiltTop, k=2, groups=cellTypes, ret.val=TRUE, text.label=FALSE, title=\"\")\n",
    "dev.off()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "From the above figure it was clear that MDS is seperating the cells based on batch. So there is a large batch effect here. This could be due to information coming predominantly from one batch. So I will now see how much zeros and ones the cells in each batch have."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [],
   "source": [
    "%%R\n",
    "cdDCell <- cdFiltTop[,4:87]\n",
    "cdECell <- cdFiltTop[,88:176]\n",
    "onesDCell <- rowSums(cdDCell)\n",
    "zerosDCell <- dim(cdDCell)[2] - onesDCell\n",
    "onesECell <- rowSums(cdECell)\n",
    "zerosECell <- dim(cdECell)[2] - onesECell"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n"
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%%R\n",
    "#hist(zerosDCell, xlab=\"Peakwise zeros in D cell\", breaks=100, main=\"Histogram of zeros in D cell\")\n",
    "#pdf('hist_of_Ones_D.pdf')\n",
    "hist1 <- qplot(onesDCell, geom=\"histogram\",  binwidth = 1,fill=I(\"white\"), \n",
    "      col=I(\"black\"), \n",
    "      alpha=I(.2),) + xlab(\"Number of cells\") + ggtitle(\"Accessible peaks in Batch-1 cells\")\n",
    "\n",
    "hist1 <- hist1 + theme_light(base_size=20) +\n",
    "        theme(strip.background = element_blank(),\n",
    "              panel.border     = element_blank(),\n",
    "              plot.title = element_text(hjust = 0.5))\n",
    "print(hist1)\n",
    "#dev.off()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 266,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "png \n",
       "  2 \n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%%R\n",
    "pdf('hist_of_Ones_D.pdf')\n",
    "hist1 <- qplot(onesDCell, geom=\"histogram\",  binwidth = 1,fill=I(\"white\"), \n",
    "      col=I(\"black\"), \n",
    "      alpha=I(.2),) + xlab(\"Number of cells\") + ggtitle(\"Accessible peaks in Batch-1 cells\")\n",
    "\n",
    "hist1 <- hist1 + theme_light(base_size=20) +\n",
    "        theme(strip.background = element_blank(),\n",
    "              panel.border     = element_blank(),\n",
    "              plot.title = element_text(hjust = 0.5))\n",
    "print(hist1)\n",
    "dev.off()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n"
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%%R\n",
    "#hist(zerosDCell, xlab=\"Peakwise zeros in D cell\", breaks=100, main=\"Histogram of zeros in D cell\")\n",
    "#pdf('hist_of_Ones_E.pdf')\n",
    "hist1 <- qplot(onesECell, geom=\"histogram\",  binwidth = 1,fill=I(\"white\"), \n",
    "      col=I(\"black\"), \n",
    "      alpha=I(.2),) + xlab(\"Number of cells\") + ggtitle(\"Accessible peaks in Batch-2 cells\") + ylim(0,5000)\n",
    "\n",
    "hist1 <- hist1 + theme_light(base_size=20) +\n",
    "        theme(strip.background = element_blank(),\n",
    "              panel.border     = element_blank(),\n",
    "              plot.title = element_text(hjust = 0.5))\n",
    "print(hist1)\n",
    "#dev.off()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 267,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "png \n",
       "  2 \n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%%R\n",
    "pdf('hist_of_Ones_E.pdf')\n",
    "hist1 <- qplot(onesECell, geom=\"histogram\",  binwidth = 1,fill=I(\"white\"), \n",
    "      col=I(\"black\"), \n",
    "      alpha=I(.2),) + xlab(\"Number of cells\") + ggtitle(\"Accessible peaks in Batch-2 cells\") + ylim(0,5000)\n",
    "\n",
    "hist1 <- hist1 + theme_light(base_size=20) +\n",
    "        theme(strip.background = element_blank(),\n",
    "              panel.border     = element_blank(),\n",
    "              plot.title = element_text(hjust = 0.5))\n",
    "print(hist1)\n",
    "dev.off()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "So it is clear that Batch-1 is having more zeros than Batch-2 and OE19. So we would be applying an additional filtering"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Additional Filtering\n",
    "We need an additional filtering on top of the `clean.count.peaks` so that number of ones in each batch is similar. In this filtering we would only keep the peaks that are observed in more than 2 peaks in each batch. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[1] 37071   176\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%%R\n",
    "cdFiltTopSel <- cdFiltTop[rowSums(cdFiltTop[,4:87])>2,]\n",
    "cdFiltTopSel <- cdFiltTopSel[rowSums(cdFiltTopSel[,88:176])>2,]\n",
    "print(dim(cdFiltTopSel))\n",
    "#write.csv(cdFiltTopSel, 'cdFiltTopSel.csv', sep='\\t')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n"
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%%R -w 450 -h 450\n",
    "MDSClust <- plotMDS(cdFiltTopSel, 3, groups=cellTypes, ret.val=TRUE, text.label=FALSE, title=\"\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "png \n",
       "  2 \n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%%R\n",
    "pdf('MDS_updated_after_Batch_Correction.pdf')\n",
    "MDSClust <- plotMDS(cdFiltTopSel, 3, groups=cellTypes, ret.val=TRUE, text.label=FALSE, title=\"\")\n",
    "dev.off()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "So this filtering clearly removes the batch effect and the cells are no longer seperated with batches."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [],
   "source": [
    "%%R\n",
    "write.table(cdFiltTopSel[,1:3], '/home/mqbsxsm2/Rna-seq_Data-Analysis/syed_scATAC_ReDoing/All_Filtered_BAM/MergedFilt_ReDo.txt', row.names=FALSE, sep='\\t',quote=FALSE)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## K-medoid clustering\n",
    "\n",
    "I would now apply k-medoid clustering to cluster the cells. As I already know that there are three cell-lines mixed I would choose k as three. While calculating k-medoid we passed the Jaccard distance as data value instead of actual data."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [],
   "source": [
    "%%R\n",
    "pamx <- pam(as.dist(getJaccardDist(cdFiltTopSel[,4:dim(cdFiltTopSel)[2]])), k=3, diss=TRUE)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "So the k-medoid clusters the cells as-"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "\n",
       " 1  2  3 \n",
       "41 80 52 \n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%%R\n",
    "table(pamx$clustering)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [],
   "source": [
    "%%R\n",
    "clusterInfo <- data.frame(CellName = names(pamx$clustering), Cluster=pamx$clustering)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n"
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%%R\n",
    "CellClust <- paste0('Cluster-',as.character(pamx$clustering))\n",
    "MDSClust <- plotMDSClust(cdFiltTopSel, 4, Clusters=CellClust, ret.val=TRUE, text.label=FALSE, title=\"\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "png \n",
       "  2 \n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%%R\n",
    "pdf('MDS_updated_Cluster.pdf')\n",
    "MDSClust <- plotMDSClust(cdFiltTopSel, 4, Clusters=CellClust, ret.val=TRUE, text.label=FALSE, title=\"\")\n",
    "dev.off()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n"
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%%R\n",
    "set.seed(100)\n",
    "#pdf('tSNE-All-Batch.pdf')\n",
    "tsNE_out_2PCs <- plot_tSNE(cdFiltTop, 14, groups=CellClust, perplexity=30, \n",
    "                              cellName=cellNames,ret.val=TRUE, text.label=F)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n"
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%%R\n",
    "set.seed(100)\n",
    "#pdf('tSNE-All-Batch.pdf')\n",
    "tsNE_out_2PCs <- plot_tSNE(cdFiltTop, 6, groups=CellClust, perplexity=30, \n",
    "                              cellName=cellNames,ret.val=TRUE, text.label=F)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "    CellName clusterInfo\n",
       "2        D02   Cluster-2\n",
       "3        D03   Cluster-2\n",
       "5        D05   Cluster-2\n",
       "6        D06   Cluster-2\n",
       "7        D07   Cluster-2\n",
       "8        D08   Cluster-2\n",
       "9        D09   Cluster-2\n",
       "11       D11   Cluster-2\n",
       "12       D12   Cluster-2\n",
       "13       D13   Cluster-2\n",
       "14       D14   Cluster-2\n",
       "16       D16   Cluster-2\n",
       "17       D17   Cluster-2\n",
       "18       D18   Cluster-2\n",
       "19       D19   Cluster-2\n",
       "21       D21   Cluster-2\n",
       "22       D23   Cluster-2\n",
       "24       D27   Cluster-2\n",
       "30       D33   Cluster-2\n",
       "31       D34   Cluster-2\n",
       "33       D36   Cluster-2\n",
       "35       D38   Cluster-2\n",
       "36       D39   Cluster-2\n",
       "37       D40   Cluster-2\n",
       "44       D47   Cluster-2\n",
       "46       D50   Cluster-2\n",
       "48       D54   Cluster-2\n",
       "51       D57   Cluster-2\n",
       "53       D59   Cluster-2\n",
       "56       D62   Cluster-2\n",
       "58       D64   Cluster-2\n",
       "60       D66   Cluster-2\n",
       "62       D69   Cluster-2\n",
       "63       D70   Cluster-2\n",
       "64       D72   Cluster-2\n",
       "66       D75   Cluster-2\n",
       "67       D76   Cluster-2\n",
       "69       D79   Cluster-2\n",
       "70       D80   Cluster-2\n",
       "71       D81   Cluster-2\n",
       "72       D82   Cluster-2\n",
       "73       D83   Cluster-2\n",
       "74       D84   Cluster-2\n",
       "75       D86   Cluster-2\n",
       "80       D91   Cluster-2\n",
       "81       D92   Cluster-2\n",
       "82       D94   Cluster-2\n",
       "84       D96   Cluster-2\n",
       "85       E01   Cluster-2\n",
       "87       E03   Cluster-2\n",
       "91       E07   Cluster-2\n",
       "92       E08   Cluster-2\n",
       "100      E16   Cluster-2\n",
       "102      E18   Cluster-2\n",
       "103      E19   Cluster-2\n",
       "105      E21   Cluster-2\n",
       "112      E28   Cluster-2\n",
       "115      E31   Cluster-2\n",
       "116      E32   Cluster-2\n",
       "117      E33   Cluster-2\n",
       "118      E34   Cluster-2\n",
       "121      E37   Cluster-2\n",
       "126      E42   Cluster-2\n",
       "127      E43   Cluster-2\n",
       "131      E47   Cluster-2\n",
       "135      E52   Cluster-2\n",
       "138      E55   Cluster-2\n",
       "141      E58   Cluster-2\n",
       "142      E59   Cluster-2\n",
       "143      E60   Cluster-2\n",
       "150      E69   Cluster-2\n",
       "153      E72   Cluster-2\n",
       "154      E73   Cluster-2\n",
       "158      E78   Cluster-2\n",
       "163      E83   Cluster-2\n",
       "166      E89   Cluster-2\n",
       "167      E90   Cluster-2\n",
       "168      E91   Cluster-2\n",
       "171      E94   Cluster-2\n",
       "173      E96   Cluster-2\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%%R\n",
    "cellInfoWithClust <- data.frame(CellName = colnames(cdFiltTopSel[,4:dim(cdFiltTopSel)[2]]),\n",
    "                               clusterInfo = CellClust)\n",
    "cellInfoWithClust[cellInfoWithClust$clusterInfo==\"Cluster-2\",]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAeAAAAHgCAIAAADytinCAAAgAElEQVR4nOydd1wT5x/Hv5fLJmHvoShuRUFERdwTcU+sde89altXW61i3Xtr6/ZXW7G17okKiltUnKioyN4jITv3++MwhksYQsgl8Lz/8HXPN8/d8z25fPLcM75fjCAIQCAQCITpwaDbAQQCgUDoBwk0AoFAmChIoBEIBMJEQQKNQCAQJgoSaAQCgTBRkEAjEAiEiYIEGlEScrl806ZNrVu3trGxcXR0DAgI2L59u0Kh0FQYO3YshmErVqzQPVcoFI4dO5Y8PnjwIFYMffr0KYdjAQEBPXv2LN9NIRDmApNuBxC0cevWrStXrvz00084juutoFKpunfvfuPGjU6dOs2YMUMul9+6dWvGjBknTpy4cuUKg/Hl133FihUjRoyoWbNmyS2GhITUrVuXYqxfv34Fb6RkSr1NBMJ0IRDVlVWrVgGAVCotrsLu3bsBYO3atbpn7du3jyyOGTMGx3EWizVgwADK6QKBYMyYMeTxgQMHAODMmTOGcr5169ZBQUFlqVnqbSIQJgsa4kAUy82bN5lM5syZM7WN8+bN43K5t27d0lh4PN7s2bP//fffCxcuGN1HBKIqgwS6mtKpU6cFCxYAAJfLpUiwBqlUqlQqU1JStI1MJvPVq1fLly/XNi5ZssTV1XXWrFlyudwg7qlUKgzDjh49eujQoTZt2lhbW7dv3/7s2bPF1b9161b37t3t7e1r1KgxaNCgd+/ekfay3CYCYbIgga6mrF+/fvz48QBw7ty5GTNm6K0THBwMAC1btgwNDU1ISNDYa9as6eLiol1TIBCsW7fuzZs369atM6CT+/fvnzp1qre398SJEz99+tSnT59jx47pVjt9+nSHDh0+fPgwderUoUOHRkZG+vr6vnz5soy3iUCYLBiBgiVVV1avXr1gwQKpVMrhcPRWIAhi+fLla9euFYlEANCwYcNu3br16dOnS5cuGIaRdcaOHRsWFpafnw8AnTp1unfv3suXL2vUqAEAQqFw8ODB+/fvB4CDBw+OGTNm8ODBderUobTSu3fvwMBAilGlUjGZTAC4ceNG+/btASArK6tly5ZqtTo2NpbJZAYEBFhbW58/f16lUjVq1EilUj169MjS0hIA4uLimjRp0rNnzxMnTpTlNhEIkwWt4kAAAKhUqpycHE3RxsaGwWBgGPbLL7/Mmzfv6tWrV65cuXz58pYtW7Zs2eLj43Py5EndNRvbtm3z8fGZO3cuqYy6hIWF6RqdnZ11BZqkS5cupDoDgK2t7dy5c2fMmPHs2TMfHx9NnTdv3sTGxm7YsIFUZwCoXbt2SEjIsWPHCILQ/JAgEOYIGuJAAAC8evXKXovk5GTNRxYWFn379t2yZcvLly8fPnw4ePDgx48ff/vtt7oXady48cyZM//5559Lly7pbUXvKo7Zs2cX51WzZs20i6Qux8XFaRvJYqNGjbSNjRo1kkqlSUlJpd85AmHCIIFGAADUqlXruhYODg55eXlDhgyhjPk2b978+PHj/fr1u3XrVkZGhu51li5d6uzsPHPmTIPMFpKjHBrIhcyUK5NjdJSeMrlGW3tDDQJhjqAhDgQAAJ/P79Chg7aFzWaHh4fL5fJhw4ZRKru7u4OOJpJYWlquXbt25MiR69evr7hX5ESfhqdPnwIAZRTby8sLAF68eNG9e3eN8fnz52w2mxwKRyDMF9SDru6o1eriPurRo8epU6cOHTqkbfz48eOJEycaNWpkZ2en96wRI0a0a9cuNDRUKpVW0Ldz5849evSIPM7Ozl63bp27u7u3t7d2nbp169arV2/btm15eXmk5e3bt8eOHevRo4f2XscSbhOBMFlQD7r6Qq5qWLVqVVBQUEBAgG6Fbdu2RUVFjR49esuWLc2bNxcKhR8+fDh37hyGYadPny7hytu2bWvevLlKpaLYDx8+fOfOHd36S5YsoYxmkLi5uXXs2HHUqFF8Pj8sLOzDhw+HDx+mLMbAcXzt2rUDBw708/MbOnRoQUHBoUOHmEwmuYGwLLeJQJguxt24iDAhPnz40LFjRz6fv2jRouLq5OfnL1++3N/f39bW1tLSslmzZtOnT4+Pj9dUGDNmjEAg0D2RnPqjbPUuDt192EqlEgBCQ0N3797t6+traWkZGBh4+vRpTQXKVu+IiIjOnTvb2dm5u7sPHDjw7du3X3WbCIRpgtZBI0wRch10aGjo4sWL6fYFgaANNAaNQCAQJgoSaAQCgTBRkEAjEAiEiYLGoBEIBMJEQT1oBAKBMFGQQCMQCISJggQagUAgTBQk0AgEAmGiIIFGIBAIE6XaCbRxVq0YpxXjBAAy2v9YVVpQhB4zhEGoXgItl8uzs7ON0JBYLDZCK2lpaUZoxTj3kpOTY6iEsyUgl8tlMpkRWkGPGcIgVC+BRiAQCDMCCTQCgUCYKEigEQgEwkRBAo1AIBAmChJoBAKBMFGQQCMQCISJggQagUAgTBQk0AgEAmGiIIFGIBAIEwUJNAKBQJgoSKARCATCREECjUAgECYKEmgEAoEwUZBAIxAIhInCpNuBagpBQMTrtHep+bUdBe0bODIwjG6PEAiEyYEEmh5CT8acfZxEHnf3dvl1UFMk0QgEgkKlD3H88MMPiYmJmuK7d+/mzp07fPjwzZs3K5XKym7dNPmQLtaoMwBcikl+lZxHoz8IBMI0qUSBfvr06ZYtW16/fq2xqNXq0NDQXr16bd++PTMz8/jx45XXuimTni+lWvKoFgQCgahEgY6Li+NwOFwuV2N59uwZn8/v2rWrjY1NSEhIRERE5bVuytRztqRYGrlZ0eIJAoEwZSpxDLp///4AcPfuXY0lOTnZ09OTPK5Ro0ZKSgpBEFj1G3y14rOW2aT9ku1IFkMHNrEXcuh1CYFAmCBGnSQUi8U8Ho885vP5KpVKKpVqLDExMWQSTBzHPTw8KsMBtVqtUqmMkJxUpVKJRKLiPiX+O9V47fLjGCObb2UtyWMSo1OcppajFZlMlpKSUgE3y0TJ92IoFAqFRCJhMCp3UoRMUG2EVkzhMTMUZX/McBx3cHCobH+qFUYVaAsLC4lEQh5LJBIcx7UHQLy9vSvbAblcLhKJbG1tK7shkUgkEAiK+zTrwQMJAINQ24mzAYDz/oOds3M5WklJSXEu14lfRcn3Yiiys7P5fD6HU7lvEnK5nCAII7RiCo+ZoTDOY4bQi1E3qjg7O3/69Ik8TkhIcHR0rIbjGwAAbHaRYvX8T0AgEKVhVIH29vbOz8+/d++eXC4/efJkx44djdm66WARMlS7yA3qQZcnCATClDGqQDMYjMWLFx89enTcuHE8Hm/IkCHGbN10ULyO1S4ya3nS4wcCgTBtKn0Met++fdrFOnXqbN68ubIbNXFkEZHaRfGe3zkBAXQ5U8WQyFU8Nk63FwiEYUBbvWlAGh6uXZQ/fkyXJ1WJ6y9TFxx7DAD+te0mdPJqVsOGbo8QiIqCotnRANvPT7uoSksj8it9sVQVRnL6dFKbdnW7tlh+ZnWL+Cf34zIn/3FPqlDR7RcCUVGQQNMAr28figWz4NPiSRVA8t+prCnTiI8fAKBRypuFl7a65qYCwNvUfJo9QyAqDBJoGuAPLTI7avXTYqjkrRNVGPHhoxRL06SXAGBjgTZnIsweNAZNAwxLS9dXL8XHj6szMjjt2qIZwoogux1FsYhZ3J7NXN1seLT4g0AYECTQ9IAJBYJxY+n2oipgMfwb8f/+1Lb0mDW8c8s6dPmDQBgQ9GaNMG8sFy/i9u5NHrN9fV2iH3ZtVRdlqEFUDVAPGmHeMKyt7XbvJBRbMBxHQ/mIKgYSaERVAGOx6HYBgTA8qMeBQCAQJgoSaAQCgTBRkEAjEAiEiYIEGoFAIEwUJNAIBAJhoiCBRiAQCBMFLbOjA4IoOHFCeukyMJkWI0dyAlrT7RACgTBFkEDTgOj333OXLiOPJf+dsg87jjQagUDogoY4aEAWcVO7WPDX33R5gkAgTBkk0DRAyahCyGV0eYJAIEwZJNA0IJg2VbvI6xlElycIBMKUQWPQNGD5w/cYhuVv38Ht1JHbowevDzXBCgKBQAASaFrA2GzLRQstFy2k2xEEAmHSoCEOBAKBMFGQQCMQCISJgoY4EMaCIAr++Vd26xbu6GgxZjTu7Ey3QwiEqYMEmgYIqVR84KD8yROmp6dgwniGnR3dHhmD3NAVol27yeP8rducox/hjg70uoRAmDhIoGkge9Zsydlz5LH84SO7A/swPp9elyodpVKjziSS/04KJk6kyx0EwixAY9DGRpWQoFFnAJDdvCWLiKTRH+NAyKibcdS5ebR4gkCYEUigjY1aJC5SxkAtEtHki/HALCy43btpW7idOtLkCwJhNiCBNjasOl6sJk20LZzANnQ5Y0ysV4TyevUij202bWD7+dHrDwJh+qAxaGNDqNWKZ8+0Leq8fNzFhS5/jAbu6mq7ZxfdXiAQ5gTqQRsb5du3FIv84UNaPEEgECYOEmhjgzs5US1oRTACgdAHEmhjw7Cz015exu3ejdu+HY3+IBAIkwWNQdOA1dJfuJ06yh8/Znp68noFAxP9FRAIhB6QNNADp0N7Tof2dHuBQCBMGjTEgUAgECaKKfagCYIQVc7eDaVSKZfL8/PzK+Pi2shkMoIgKrsVpVJZZe5FLpdr/q08lEqlcVqpno8ZhmECgaCy/alWmKJAYxgmFAor48pyuZwgiEq6uDbGeVLFYnGVuRelUsnn8zkcTqW2Qj4AxmmlyvxpjPOYIfSChjgQCATCREECjUAgECYKEmgEAoEwUZBAIxAIhIliipOE5o46N1cdH080bIihHSgEIfnvlOz2HYajg2DUKIaDPd0OIRDmRLVXEEOTt2Zt/uYtAJAP4PDfv+wWLej2iE5yl4eKdu8hj/M3bHSJecKwtaXXJQTCjEBDHIZEdvMmqc4k6f0G0OgM7RBKpUadSQr+PUmXMwiEOYJ60IZE/iiaYskYGoJxOOzWrYVTJgOO0+IVXRAFBRSL7No1wfhxtDhTDiRy1fbLsWH34r0bprNsniRIXrR1aze2yXgHHsp1izASSKANCe7mRrHIbkUBgDT8GqhUwlkz6XCKNhiWlnS7UCE2X3x18kECk5eZyv0TJAAANxMjJUrJkoBf6XYNUV1AQxyGhNenN7dzJ70fyR9Ux6j8nIDW2kXzGpE/+SABANjCj9rGh6kPcmQ5NHmEqHYggTYkGJttu3+f7Z5dnKlTcQ/3Ip9VfswEE0Q4b57mmNOunYX5jG9oINRsioWLV+5OcQRCAxriMDAYk8lq0kQ2aQrFzqmWUfk5Aa1dHkfLoqIwoZDTvp15rTtsVcf+7tsMWU4doUe4xtjdM4jL5NHoFaJaYU5fGHOh4NhfFIv12jUW3wyjxRnaYTjY8/r1pduL8mDFZwKAWsnNeD6e7xDNYIq/69ytW83udPuFqEYggTY86rw8isViWAhgGC3OIMqNu40FeaCWW4gS2wbUtQ/y9KPXJUR1A41BGx5up47aRf6A/sBA/8/mx7CAmi1qkdtqCAAYEViLXn8Q1RDUgzY83K5duYsWSn9bCQD8Af2tfl1Kt0eI8mDJY20Y4Rf1Jl2mUPvVsrUXorlBhLFBAl0pcEaPsps+jW4vEBWFzWR0bOhEtxeI6gt69UYgEAgTBQk0AoFAmChIoMsEoVTS7QICgah2IIEuBWVcXOaoMUk1ayW6eRT8eYxudxAIRDUCCXQp5Pz0s/TqVfI4+/sf5NHUeHUIBAJRSSCBLgl1VpbsRoS2RRZ5ky5nKg/lx49ZEyclunlkfjtCGh5e+gkIBMIoIIEuCUwopFgYNja0eFKJEETO/IWSc+cBQHr9RubI0cq4OLp9QiAQAEigSwZjsQRTi4Q9MtOwEiWgSkyURUZqW6rkWwICYY6gjSqlYLV4EdvXV3b7Nu7iYvHtcIalpfLjR/Ef+9RZWZy2bflDh5j7Nm5MIKBaLCxo8QSBQFBAAl0aGMbrFczrFUyWVImJqW3akscF/55UvH5tteQX+pwzAAxra/7gQQVhJzQWbvduNPoDAKrkZEXMM9zdjdWoEb2eIBD0ggT665D8d0q7KNqz1+qnxeabbFAV/4lQKqzXrGb7+MgfPsTd3S3GjaM3VVVBWFj27LnkscXwb6zXrEaBABHVFiTQX4cyPp5iIeRyjGd+EdwJiSR75izJ+QsAwGnfzmbDeouxY2j2SanMXbdetGOnxiD+35/c4GBKdEAEovpg3uOnRkZy8j/x4SPaFl6vYHNUZwDI37GTVGcAkEVE5v22kl5/AKBg9x5tdSZRvnlDizMIhCmABLpMqLOyZFFRWdNnaBvZzZpar1hBl0sVRPHsmXax4J9/6fJEg/KRnk1ArAYNjO8JAmEioCGO0pGcOZM1eaqune3nx3CwN74/BgF3ddUu0j4xCACAU7sLFqNGVs9cjggECepBlwKhVOpVZwBg+foa2RkDIiwarlo4YzpdnmjgDBigXbRe9Zv1yt/ocgaBMAVQD7oUVJ8+6bVbjBrJH9DfyM4YENzV1fXVS8mlSyCXc7p0xh0d6fYIOD2DOH/+r+DEP8DA+EOHcAIC6PYIgaAZJND6UaWmio8cJXJz2W0DKR9Zr17JCwpi2Jvr4IYGTCjgDxpItxdF4LRvh8Y0EAgNSKD1oEpJSfHzLyz8sY8bHCw9d44s8YcMsfj2W7Qy1/gQcnnuL0vFhw8DgGD8OMtffsaY6OlFVHHQI66Hgr+Paxel5845346SP32Cu3uwfZrR5VU1J3/jJlKdAUD0xz6Gvb1w1kx6XUIgKhtjC/SSJUueP39OHvv4+Pz0009GdqAsqLOyKRaGqwuvhgctzlBRq6XXrqlSUjn+/mBJDbZXhVE8f6FdlD94SJcntEAAcS0+/HZyFJPB7FazR3PH5nR7hDAGxhboxMTEJUuW2NraAgCHY3J57AmxWHTgoOLVS20jL7inkd+mlR8/5m/YWBB2gh8y1HLuXNzDHZRK6bVrqrR0yalTspu3yGqMRQuh8nOHK29FiRIT2Y0bs1u1rPjVpJcvFxw/AYSa16cPr2+frziTWWQ/PcatrIeHICAlWwxqtYu9Cf3+nY07s+fpLvL4VuLNpW2WI42uDhhVd1QqVU5OTuPGjRkmGQGOUCqzJk+RXruubeQF97Ra9qtR3Sgo+BKP6a+/VYlJtnt2Zc+YKQ2/BgBAAHweAFf/thImT4LK/PHInvudZsBHMHGC1dIlFbma9MLFzPETyGPJufM2BQX8YSFlPNdixLfSi5cAgPwvsBgzuiKeFEd+viRi3ByfqHMAcKlFl4D9W4S2dEYm0fAorcgbQ3j8FSTQ1QGjCmV6ejqbzf7tt98mTZq0bt26zMxMY7ZeKopH0RR1drhwznbvHtzFxZhuULJqyW7eFG3bXqjO8EWdSVQZGZXoyZOn2sPxor2/q+L1LzosIwUnTmgXNXvNywK3c2eHM6cE48cJJk50vHCO06ZNRTwpjsiFq0h1BoDGD67e/85EQxWq1Cq6XUAYA6P2oHNycmxsbIKCglxcXI4fP7569eo1a9ZoPo2JiRGLxQCA47iHR6UM+KrVapVKJZfL9X5KfPxAsWTGx2MODuVoSKVSiUSicpwIAERODsUi1onQpCGdICAlpXwNle5J0e3gAJD29AnGZpX7guqCAu2iTCJJ+ey8QqGQSCSlvFq5uMC0qQAgASjfXavVagAooRX+i6faReaHuJSvb6jkx6x8ODOdtYsN+A1TUlIq8piVHZlMVsb/BBzHHcr1fUEUh1EFukGDBtu3byePp06dGhISkpeXZ/k5uKW3t3dlOyCXy0UiETkCrou6S5fkohbnDh1049mXBZFIJCjXiQBAdOmS2bat7GZhWhNOu3b8Xr2yz5zVrcnYvdO5Mnv3yrZtU4tanAIDK5L0SxwcnHP9hqYo7NBe6FyoO9nZ2Xw+v7KnJeRyOUEQeltRJSVJr161lOZpG3EO29nZWbdyqa2U8JiVj/GOE+S47EnGEwCsk0enfg36Q8Ues7KTkpJSjv8EhEEwqkDHxsbK5fImTZoAAJPJZDAYTFNaysqwtbU/EZa/ZYvsRgS3cyfhvO/Kp84VBONybTasy9uwUZ2RiTs6CL+bizs7y+/cER88RFawO7Afd3Vh1qmTmk1dbWJYmDVrWK9ambNgIVm03bG9gikZLYZ/QxQUyG7cAAxj+/sLp0w2hJsGQP7wYXrf/gDgVNTuNnksLf7o8jDt4aWP5BA8HHv1p59ji/q2KIxU1ceo+iiTyVauXPnzzz/XrFkzLCysUaNGfD7fmA6UCqd1K07ro3R7Abibm836ddoW699WCGfNVGdkMOvVw9hso3liMXIE0b0bJyeXWcOj1MCqsohIeUwMy8uL272b/kxgGCaYOEEwcUKl+FoB8rfv0C4W1KqT59u61oghbq1MZSLu0oeL2sXTcaeQQFcHjCrQ3t7eISEhq1evlkqlTZs2nTdvnjFbN3dwZ2ecljdNCwuWk1OptXJ/WSL6Yx95zOvd23bndjPK1vh5fUghtl6edbfSHyBbGzWotYtSpRQAMqUZz/Jj7HkOdazr0OQXonIx9ghDv379+vXrZ+RGEZWNKjlZo84AIDlzRjZ2DKd1q8poi5DLpRcvqnNyOG3bMmvVMsg1+SFDC/76W1NkNWlskMsaEHeB+wO4ryk2tm9y/dO1DQ/XkasuO3t0me03FwMUgaCqYUJDwAi9yG7dkkc/ZtaswevZs1KXPFcEVXIy1ZKUZMDrqzMycn9bWfDX39z2HVSZGYpnz0ktst2zi9erV8Wvb7VwoTotjVxkyevVyxSCr1IY3nCESC66En8ZAPp49etVu/fgUwPg86rL8E9X27gFtnSulF9EBI2Y6BceQZL320rN8Ci3eze7vXtMU6NZ9epRLGxfn3JfTfHqVf6GTZKzZ3n9+lr+8D2zVq3sH+eToxDSiBsAXxaDZ02a4pZYoaXZJAwHe7vDh1SfEoDDxsswnmN8uDh3VvM5M3xnYRiGAZZWkEap8Cn/ExLoqofZjBJWQ9Q5OdqTV9JLl6U3bpRQn0YwgcDu4H5N0XrtmnIPPqjz89O6dJOcPQsAkv9O5SxYpMrIoIwRF0GpLF9DVDAMr+FhmuqsgYExyHEMex412m1Ny5p0eISoXEyxO4YgUaWkUi1J1JEE04Hbtavrx/eqxETcxaUi60zk9x9oF2U3bypfvy620c6dTfOVorJhYIz5LReuvlc4k9nDM8jPqQW9LiEqg+r4cJsLTK/aFAu7hZ8hG1AqDatuGJPJrPkV/ThVaqr00iUAjBcUpImFpLtTERMIKZN4JJzAQMtFCyrisFkT6Np2V/u9qcoUe56Dh9A0Qi0iDE31Fmi1WvHsOaFUsJo0Mebi4jKCsVj2fx/LGDqMLFqvCGU1bGiQK0vOnc+aOAkAuJ07WS6Yz2pMw6IFxbNnaT16ksc5CxZyT/4LTb0BgN2iBUdrIyW3Ywd240aslb+x6tWT3bmLOzkKJk3EhEIiN5dZq1b17D5rsGRbutq6ll4PYbZgBEHQ7YPx0N6Dq87Pz546jZy45wQGWq9ZxfT0NFRDhtyDq1KpkpIZTo66PyHl24OrfP8+tW17bYvbx/clKJ1IJOJLpQBg2CxfWZOmkAPNJHhwT5ttW8lN2Kq0NNG27cr3H5j16gqnT2MYbs90CVu9DUhlbPXWC/mYJYoS/3x1NCLhRmePLt80/NaJb+BhdLTVm0aqbwdEtHWbJnad7NatvFVrbHftKPEMmsBx3MPdgNeT371HsSjevmM1qK+3MiESFcyYmXv5CgBwg3rYbNrIEOqJkqyMi1O8es308mLVpy7nKA5CIilSzv8S9Ad3dDRyiFfzRaqUTL0yiTwO/3Q1U5r5c+slbNzkXgcR5aP6ruJQxL7RLkpOn6bLEyOjG0+DYVtshI289RsUl6+Qx9ILF/PXrNWtk79te2q7DlkTJ6V17pK74rcyukH51cG+5vVFnZeX99vKzFFjcuYvVH1KKPuJVY/X2a8BvrwEP0l//DbnTQn1EeZF9RVopmeR6SxezyC6PDEynA7tOe2+ZM7mDx2COzoWV1n59l2R4oePlAqqlJS8las0RdGOnYpXr8rihuX8H7ldu5LH3B7d2WVPDUMQ2TNm5W/fIb16VXzkSErrAHVeXulnmSeXPl7se7JX35O91j1Yky3VExuLgTEoMcIZGK5bDWGmVN8hDuGM6YrnL2RRUYXFed/R608ZUb57l791e8Hx4wCA9emtXrtG75hDCWBcrt3ve8RHjio/fmQ39+UPHFhSZWGRkXTdeEnK9++pljdvWA1Kj+PDsLKyO7hflZQEOI47OWV/jsynTs/I27BBlZjEsLERzpmlu55aGR8vvXpV2yKLjDTIfkJTIybj6bboLeRxRMINFaGa77+QUqe+TYNmDj5P0h+TxWYOPiguR1Wi+gq0/NlzTtcunE4dmLVqc9oGfq3M0ULur8tEe/ZqisTpM3lWVtarV5Vwil4wgUBQtlCfwpkzJP+d+lLU2QPN9KLKAatBA8XTmIL//gMM4w8cwGrUqITr465FFyGoVFmzZ8tuRJClgrAwl5fPGZZFk07pBMIn5IpSb8QceZhaZEn4rcSbSj8lk1HkO8vG2bObzzn2+limJMOJ7xRSfxilAsKsqWZ/S5WKSE0FS8vsH378kmpv/DizGN+QP3igrc4k4iNHyyHQZYfVsKEw4jrjRgQA8HoG6Wb/wh0drJb9mvtLYa5Cy+/nqZJTMr4ZThZFO3fZnwgre9Qk5bs4jTqTyG5FUf46zNq1Oe3bySIiNRZu27Zfc09mA59pQbHgDD3DF/Y8hxk+M43iEcLYVKMxaOnly+l16km7dE+sWatIqr0/9ik/FptTynRQvI7VNXI7d6rsdhlOToJxYwXjxhaXm1Ewfpzzw/t2R484370tnDtHdPCg9qfiffv0nqUXQq2mmnQtOG6zbi1/8GAA4IQthScAACAASURBVPUKdjx/juFgyPV/pkOnGp21iyH1h6F4ddWN6tKDVufmZo4ZB0BNukqiSvjErFnDyC59Lay6dXWN3O7dje+JLkViVRcdcCCksrJfh1W3DqdDe+1ONCdQT3JY3M3NZvNGm80by+Or+eDAc/i9+76zcWfy5HnNHf3auhdO7aoIVWRCRLI4yZXj1l7QAal2Faa6CLTixcsSPjXUDr1Khd3S32LMaPGBwv4pp6W/YsQIi0FFpvgkZ84U/PU3YAxWo4bCObMxLrd8bamzs0U7dyli37AaNsBGjoCv2XTDql9PGh7+xe2iYe0IkShn6a8Ffx4DAKufFgumTilyMo7bbNyQv269KiWFYWUlnD2LYW391c4TxIUnyS8Sc91seP383PkcM37IHflOY5uM17YQBLHm3srbybfJYkxuDBrfqMKY8bP7VeAuxW6Fst2x3YB71SoV6xWhFiO+VX2MZzVpjLu7U3ItyyIjsyZPJY+lV68SYrHV8mXlaIVQKLKmTScHeaWXL7OiowVHDmM6Ww1Vycnq7GxmnTraWxzl9+7n79ylKVqMHCEoOq+Yu3wFqc4AkBu6Andx4fUvksABd3KyXrsGSiOnQM5j4RyWnjHZlaeen36USB5HvclYN9xXbzVzQfHsGflLSXYjXmQ916gzAFz6cGFIvaEG3z2IMBGqyxg009OTE9CaYrTdtcv13Rtev760uFQ+WA0bcoN64O569hZKzp7XLor2Fcb/lF69mvntiEQ3j6yp01QJpW/rUEQ/1p6CU0TeVMTEFKlBEDnzF6S0aJnWrUdSLS95dLTmk/QBRXr0bF9fjFUk+JH4yBHtImXBXFlIzZXOOfwwaPW1DqFX1p97qS4aqyBbLNeoMwDcj8u8/Tbja5swHXJDV6T16Jk9c1Za1+55q9cAQJY0i1InS5JJh2sIY1BdBBoAhHNmUyycNq3LPQhAPzqzZxhLz/uQ8s2bzFFjpNdvAIDk1Omc+Trh3whCeuFi3voNkpP/gUoFupuwdSySU6fFR76k1k3vXfgLRyioy90Uz59TLJRZTUredN2mddly6fWdz5p7/G78GS05BoA8CdWHHLG5LsJTvn0r0nodyd+yVfnxY30b6qZ8T6vCpeKqpKTsed9njhydPec7s5j3RpRKNRJoTps2HK3tDMJZMxl2djT6U27UOTnZM2YmetRU+fnnb/gyUcYbOEC7mnDWTACQavWFAUB6/Yaq6MBI9g8/Zo6fkL9hY9b0GVlTp4FazWruS2mR3ayZdlH+9CmlAlFQAACUzjIAML28KBZuly7aRYsR35IHqvBrGbXrJNWplzFsuG60EG2uPivi/+P4IvvrPGypeeKb1yp2I7uJo2cT0Ls4R77Td37fayy/tlnOY/IAgJDLs+fOKzj2lzQ8vOD48dQ2gYRYLFFKYjKexuXGGdVvhOGoLmPQAAAMhuXWzbk9e/Ays9i+Pmw/g8ZWNiK5y5YX/HuSPM5bvwGv4UGuOWP7+tqHHRcfOEhIJJyOHQSjRgIAQ2d+D+N/kTDlx3jNiDAASM6ek99/wG7V0vHShbz1G6QXL/F6BuHTpmIWRRbk6gZ91lzT7uD+zNFjyWNuly78YSGUmhZjRuMuLpKLFzEez2LkCHLPoSopSTar8P1GFhmZHhnpGvuK0qiG1nXs72iNWtgLi4SmYzCwQ1Pb7LgcS9ZZMbRZDTv91zF9WPWpGzJZDRsAQEePTu3dO+TKcphKllBQuMFK8fy5JkYrSdz1/+YqCl90AlwCvvefz2JQf0ERJk51EWjl27d5a9ZJzp5ldOrI+fFHVlNvuj0qP5TQ9bIbEaRAAwAnoDVlqJ3bvZt2kT90iPbGPHWaTtKW1FQAYDVubLfvD9IiEokodfhDh0gvXdLEArTdsf1Lc127Oj96IL97j2Fvx2ndWpWers7IYHp5aQ8lcXt05/YosjpQ/ugRpQnFixdsf3/qnQMAQG9fN22BHtKSuj6ynrNw00g/hUrNws37BRGv4WG15JfcXwtneq2WL9MsRWdgDBuure6fRpvrn8Lh89T47eTbF96f7+NlTtMtCKgmAk3IZKkdCoc+1deup1277vLsqW5QN3OB26mjRhxBX3Q6bRg2Ns4P7on27FUlp7Bb+luMHKH9KasBdX0hW2d8QxeMy7U7sF8afk2dmcluE0DpUONOTry+fQAgd+mvor2/k0b7sOO6k7RfnNQZa9Ife1qtVr5915GPH5rcOuptpgWH2d3bxYqvv1dovuqcL89PFCW6WLhYcawEkyby+vdTxr1neXmVvB+H1aQJp107WWSEZqn/BZtE7THMD3nUAROE6VMtBFrxkroIWn7vPqUTZ0Zwg4O1BZqiubrgLi5WS36hGJXv30vOnAUGw2bjxuy5c0mjzbateteH6IHJpPTNKUivXtWoMwBkDB5SQvptjr8/3r69KqJwfwp/4ADdGEnq9AxNmA67zp1Hbd9KjdFRJbj08aImQNIE74l9vfrjjo7a4QZVauUfz34/E3caALq6d5/afBo5cIGxWDYb1+et36BOTWPY2wnnzGmcuPdR2pdXE5QWyxypFgKtu1TDjBdvAFgM/4ZZw0Ny/oIEwxwmTviqNIAk8nv3tdfDOZ49w3Bxxu3tATfYemHF0xiKRZWWjjs66K/NZLI3bcDPnYePH1lNm/L1LXzMXb1as8NQGh6ev36D1a9LDeWtiZAjy9GoMwD8HrPX37mli0WReFJhb8JIdQaAKwmX7Cxsv204kiziLi42674E7B5kMVQj0H5OLXrWqoIB/6o8VVmgCblc9fEj7uzCqluX27mzZnsbp21bdquW9PpWQTht23LatpWmpDDLnotIpSr4+7j05k3czk7xpkhM9/zdu213VjSbjCotTRZ5E+NyuJ07YzweXoM6NIzbFdkNpEpKEu3eo0pMZPv7W4wZjbHZ3OHflJCMSp2apl1UxlXBF/b4PGq47fe57ykC/Tq7SLjtdznvoBi87b2PBP/5IvO5gCVobNcEw9COcPOjygq0NDw8c+Ro8thqyS8227eKdu6SPXuu9nC3nTvXrHvQ5SM3dIVuMDwSQlxQ7suqc3KUr2PVGRmZkyYDABAAGLg8jub17SM5fUZ6+TJZzWbdWu3uuTo7O8W/MMSd5PwFRUwMlLbpUXvxCQCYy+bPr8LZghqOiqLOAMBmFElnpTe+nQZLtmVrlwCD+Iaghaop0IRMplFnAMj9dRknsI3l/B/JbJ5V8rtdCipVceoMAKzGJYVsLgHJ6dNZU4pmQsEAAPJ37bL6+Se7fb9Lr19Xp6axW/pTFkRTNhAW/HuS991c4FOXMGsjnDtbcuaMdrF8PpsyjnzHwfWGhsUWrtLpVbtPLSvqWHyv2r2jkm5pij09g43nH8LoVE2BVr6jvvfJox+zGjemxRlTgFAqKRZm3brKN28AgD9wgHD2rPJcUyajqvNnVPHxAAAMBrdzZ70ViAKdHYOl7SFkNWjgEvNEGn4NYzI5XTrrSbCgVitevQK1mtWwoQEH040GQRA3Eq6LFaLg2r08BDXq2tStp7NpEAC87Ztu7Ljl+qdwAgh/25bNnApjUUmVkpuJkWJlga+jbw3hV09LIEyTqinQ1DwdAEyPaj2FjXE4vF69JGfPaiw269ey6tVXvH6l/JSg/BivnY2bEInkj6IxDofdwq8EpdPNT6iB5V3KMnNO+3ZFih3aYzp/Ml0Ytrb8wYP0fqTOzs6aPoOcReS0a2ezZVMJiRZNk91Pd557f4Z8B/G2b9rJo9hI317WXl7WXqC1RD1fnv/tuWGaCj+0mN/OvX0l+4swBua6VrRkFK9ea/eXeb16cdqZcdINVVqaaPee/K3blLHlT9hsvSKUVDdOh/a227ex/fzyNm1K7zcge8bMtE5d8tatJ6spXr1Oqt8w45vh6QMHZYwYWUI+Vqa7m147L7inYNLEkp1henra/+8oGZeD17+fzZrV5byrz+StW69Z4yGLjMxbVdELGhmpUnLu/VnNEuaYjKfaIetK5cKHc9rFtQ/M7PYRxVEFe9DSa9czR4zUFIUzplnOnw8Mc/0pUsbFpbbrQB7nrVptd+Qwt1PHclyH4WBvs3mTzeZNhZeNfSPatbvwMwzyN26yCAnBPdzz16/XnCKLiBRt3ab285NIpexWLQu3salUquRkhp0dZmFhtXxZ7s+FK6z5gwYJJk3ABAKmp2dZ/OF0aM/poNXLy9aTsrrsqOKLLLJWp5tZBDuxsoAomkwiX55f9tMzJNT7JQgCLduoAlRBgS7480/touL1G/NVZwAQ7S4yuSfef6B8Ak1BGUcNoKOMi8M93CXnisQsFf99XL1jpxgAAOwO7GPY2qb37U9+ZLlgvnDmDG7HDvLHT5g1a5QQ24RQKHTjKJUDdWam8l0c7umpu56aYV9kLyImMLP4G3ZcOx+t5NwA0MyhSKKDN9mxYbF/y9UKR77T8AbDrThF8hjUt21w/v2XTnRL51ZInasGZqxcxUHN8axS0eSIYVBlpGsXyxFAWS/MBtQJKNJCydCqzvjSNcscM06jzgCQt2q1/N59Zu3a/IEDilNn6cVLiW4eSZ61M78dIX/8WG+dMiI+cjS5qU/6gIEpvs2/9P0/Y/l5M2Rhcd53FWmLFqY0m+rvXLg8f16LH7TXb2RI0ufdmHs7+fbD1Afn35/dHL2Jcm4nj869avcmj5s5+IxtMs44PiMqmyrYg+YEttEsv4WyBZcwZTj+/tILFzVFfshQg1yW6elptXhR7orfyKJ16HLcyQkAhN/Pk5y/QBpxNzdVYpFoy1D0PVz+4AG7pf6QRgCg+pSQOa4wXZP0+g3p9RuucW+x4reilIA6PUM7knXu8lBu927M2rU1FryGh+vrl9IbEQAEp317PWs8TB43gfvPrZeoCTUDo3abtHdsA8CDlPsZkgx73pfQHBhgk5tOHdFwlFghcuA7oiyFVYYqKNCC8ePU2dn5m7cAgGDKZMHMGXR7VCEsJoxXvHlTcOwvAOB272a1eJGhriyYNpU3cIDy/Qeml5dm0IDVoIHr65fyh4+Aw1F9+JA97/si5xT94jOKyfNNInvwgGJRvHrNbta0HK4q3lAzmitevCAFmlAoVEnJuKsLJhDwepn9omCKOhME8SbnTaKImgcH1xFxALBgWViwzGxsB1EyVVCggcGw/PEHyx9/oNsPw4AxmTbr11kvX0YoFAwrK2VcXO6vy9Q5ubiLCzFyBJR9q7c+imTj1rQoEBRO3/m3kEVFFZz4h7Q7nDopi7qtWSDB7dRRI4iqtPS831YWHD/O7dpVMGkCJzAQ9IXZw8u7RUg33gipzto7ZWy2beUP6E8905yRqqRr7q16kHqfYm/tEmDDrX6braolGFE0pZspQBBEyYFuy41SqZTJZBbFRII3IDKZrISwEuWGyMvLb/klaCfRqqXVvj8qe1OGNDaWVVCAN2gAXC4AqB49Uj58hLu64t27aab+CqZMU16/rjlFcPE8o2ZNQqGQTJ2mvHkLAIAAVr8+vOKX0xUUFLDZbKZOaloNsq3bZNsLA4awR43kLlpIZGbmBxZZTy24cY3hVFL6VKVSCQAltGIQDPWYnXz/7/G4vzRFW66dh4WHh6BGP8/+fCYfyvyYyVXye+l38+R5jWwaewo9v9aN3NxcKyurstTEMEzwNQngEaViij1oDMOElTOGKJfLCYKopItrU8EnVRoeLj5wCACYnjWFP3yvGVGV3ooq0srde9yUFDIpSeWB1a9f5F46dIAOHbQrqDMy8rTUGQDwO3cFTZoAgPDAfvHR/yk/fGA3a8ofOLCkbS9KJZ/Ph0fRiseP8Ro1eEE9KJWFixYq+vdTvHjJrFuXHCeRPYoushKNAM7799w6dUq4F/IBqIzfTt1WKv6YpcqLJPfKkmYeCDqkbSnLYyZVSdfcWaVZHzK12bSvDWsnFouN8JVB6MUUBbqao3gaox1IRJ2dbbN1Swn16UdH7zQzgRiPJ5gwvoyXKVi3vmDHTvKYG9TDbs9uikazGjViNfoSNgSnjIBjenaQmjWUCM6aNR5fxfVP17RX7+18siPIMxgtwjMXqtoyu8cfsy8+TU7IKn94NtqRXLqkXSz4519CLieP2W2KRiZr3YpVt67RHCsOhlBIGfzl9ujxtRch8vI06gwA0gsXpTdulHwKs15d/jdf9jfzhwypYuFW+nr1b+7YXFMcVv+bclwkQ5JOsRQozfjbUd2oOj1oNUH8+k/MxafJZHFBn8b9W5QtOYiJobunA/u80YZhaekUcT1/02Z1Ti7u6ioZ+a2JRAWyXruGWa+e/FE07uQomDK52MD8xUMUDfcMAKrEpFLPslm7hhccrHj5ktWwAbdTscErzBQek/dLwK8vMp/LVLKGtg35X7lCI60gLUeWU9uqSBxBH0dftNLDjKg6An37TYZGnQFg1ennwT6ubKb5vSLw+vTOW/MlL4bFyJGgNanF9PLSjHhIU1KoJ9MExuMJZ82s0BU8qes02H7N9VUEgoALT5NuvEzjsfGB/h7enTuRMT2qJAyM0cT+qxMcE0DsebrrbFxhdNb27h0iEm4AQFOHZqMbjTWwi4jKpOoI9KdM6otbSq6khp35dRaYtWs7XjiXv2s3IS7gtGppMb5a7ArDWCyrwwdzNTkWlv2qPdyszZ+3P2y5+Jo8Pv8kae+EVt4e1nprlgohFivevsNdXMrR5Tdl7qfc06gzAEQk3DjW+3i+PN+R54hGn82LqiPQdZ2pE82u1iUFgDdlWN7ettu30e3FVyO/fz9/126Qy3E3d8sf5unm6i4ZVmCg28f3ysRE3MmphJQ39+MytYunoxPLJ9DSK1cyRxd2J4WzZlrO/7EcFzEpnmc+Ox7796PUh24C6uCeXCV34pe0ABFhmlQdgfarZftNQM0/bxcGKV41zIeJo86C8VB+jE/v/yURrSo1xW7/vq++CpNZag7c22+KRG6TKcoTa4VQKjXqDAD5W7ZyO3UqYdu66ZMjy1kYOZ88ThQlUGLjWeECZVwcxuNRl74gTJuqI9AAMDuowdDWNVNypXWchEJulbo1AJDdvi27eQt3dOQPGoiZ3nYA2eecvCTSS5fVmZlf24kuC8MCah67/SVXQMeG5ekYFuZ80UL+9KlZC/TLrJfaRW11nlVrTOaIUbLISADg9e1js3kTxi5MbKjOyys4clSVnMxu3YoXHAxoAMTEqGoq5mLNc7Hm0e2F4RHt3JUbuoI8lpw5a7vvd5OLB8TUWU+ib4WJ5Nx52Y0Iho21xaiR5Vu2PLVLXQA4dvtj6zr27Rs4dmpUHoHW7UgydWYpv4pHqQ8/iT7VFNb0caQnOBefSX3sZ/jOxDFmY7vG7HnLJZGRpFFy6jSrSRPh9GkAQOSLssZPlEVFAQDs2y+YONFq6S/G9RpRClVNoKsmarVGnQFAFhUl+edfi9GjaPRIF263brDgSyAnXp8+DGvq0HD+lq15q9cUHm/d5nzvDu6mPy1LCXBY+JygBnOCKrR/EuPxrJYuyV36q8bb4tInloUt0ZuvfCxcvR7k2XNCo0kV8a18NLJr7OPo+zgtmiy2cmndrWYPDLAn6Y/tT5/Wrqn4HPdVcv58oToDAIBo717L+T9gvCrYvzFfzFugM/JlqblSTwcLC45530jJqMViikWVmkqLJyWAOzs7hl8Vbd+uzs5hNfUWTpuqW0ejziQFfx8Xzp1TxuurkpKkV8MxNpvbM4hhaVlxhwUTJ3DatVNER+Me7pzAwHK/3X/Ie69RZwC48OF8UI1gayjnwpJyw2KwFvgvPB13KlGUWNvKK7hWLwywu8l3VtxdPr2xZZPnX1KX4e6FU4hqnSw26txcHAm0KWHGurb32ts/rhdm7141zKd8Y5FmAUMo5HbsIL3+ZWcdJ6B1CfXpglW/ns2WzcV+rJNZXJ2TU8Yry+/f/zID+d085zu3cY8vCxXUOTmKpzEMO9uv3UnIalCfpZO44GtJL6Bu1cuQpFvzjC3QAMBnWYQU3W14Jf4yAFzuaq8t0IIpk8kDTtGNqZx27XRDGyLoxfz2cZA8T8jVqDMALDj2WK5U0+hPxSFEorzfVmaOGpM1aYo8OpryqeUvv3DaFUZus1wwX3NsTjCZvN69tQ2cjh3LeGr+9p1Fizs0x7KbN5Mbe2d8Mzyte1DW1Gm6PwOVTW1rL4qlllVtvTWNj4pQAUBsPcHi5Q2OfuN2fVYH19hX+OeAfyxvb5sNhSkouZ07Wy//lTZHEcVgrj3o18nUbNPxmeI6TiY2b/Y15CxYWPDvSfJYcvasU+QN7YwhrPr17I/9T52VxbC0hEqOlll5WC1fBgxMcuo0AFiHLi97ckXtFDkAoNLaQpkR8qXPKDl1uqBjR0MlnSkjQpZwTvPvNj3aQBbntfjBim0lkldKvFxtCIIIT7wak/OUzWAH1+7V2K6Jbp1Gto0fpNwHgCxb9s22dmMa98eKBkHlhwzlhww1VNJIhMEx16+6uy11E4pZL94gpFKNOpNIzl8gp9q1YZQ34L2JgDs6CMaPY/v5MT1rcrt0KfuJ/MGDCsJOaIrspoW7nwmplFJT8fJVxf0sIxmS9F1Pdt5LuQsAIfW+6eDR0ZHvyMbZ8s/BrSqVv2P/OvryMHkcmRixpv26BrYNKXUG1B1YoBSHxR4HgCH1QvrXGUi9CgDoC/+CMBHMdYjDv7Zdj6ZfVkrN6lHfvOcJTS9tQmWQt2Zter8BuUuWZo4emzV1OqjLOipluWhhYZIXAF6vXoLPM5C6Gw6Z9YwX3u/3mL2kOgPAX7F/fsqPZ+Nso7WuUWeS8PirubLcR2mPPuZ90BhxDB/VaMyp/mdP9T87stEo3WyHpomvry/2GSaTWbdu3VWrVqnL/LSUgFAoPHv2bMWvYzTMVdQwDH4d1LSfn3tStqSRm1VtR5PbuPFVYDwef0B/7U40Jbt2FUCTKJJEcvq07NtvyjiYjjs52R89ooz/hHHYlIksu8OHMkeOIrPZ8noG8YcOMbDfxUAQRFTSLW3Lk/THAa5tjNO6Lsni5JHnh5PHnWt0me0716zDbgwePPj7778HAIlEcvPmzZ9++onFYs2bN6+EU3r27NmiRYvly5cby0djYK4CTdLc07a555fipZjkS0+TAYPAeg79/TzM6/m0XrWS4eSkjH2DcTmCyZO0B6CrBtQE4QDKTwlfkdoEw5g1a+iauZ07uTx9LH/4iGFny27e3Gh74XTlz4pTprxQhqKvV79T7/7TFLWj8ofHX23p3KqNa6Ax/TEszs7OrVq1Io87duz44sWLsLCwkgW6SmIerzxl4erzlF/Cnt6MTb/5On316Rdh96h7eU0HQiqVRUbKbt4ktAYrMYHA6uef7A4ftN27h92iBY3uVRJMnWRUbF8fg1yZYWfH7d6N7edn5J3Kk5tO0S529zTqS8+oRmP61Ozr59TCz6nFpGZTKJ9qD3SQqBISsiZNTnTzyPhmuOTceSN5aSCEQqHt5wmYd+/e9e7d287OzsLCwt/f/9KlSwDg7+9/4cKF0NBQX19fAMjMzBw5cqSzs7ODg8PIkSMzM78E2MrKyho8eLCdnZ2bm9vGjRtpuZ2yY949aA1ypfp/tz5oW+68zRjSSk+Hi3ZUnxJSWheuP+W0aWOzdfPXLj6VnL8gPnRIFhGJdeumWrG8HJvxaAHjcu0O7teEKLJa8gurIXVSy7zoVbtPDcuaD1LuC9mWXWt2s+YYde0zG2d/U/dbMichAcSeJ7u0P/UQUh/+nAULpdeuA4AsIlIWEel45ZIp//+r1WqVSgUAUqn01q1bYWFhu3fvJj8aMWKEWq3+3//+Z2FhsXfv3pCQkMzMzPv372uGOAiC6NGjh0AgCAsLS0lJWbRo0eTJk8PCwsjT582bt379+hUrVuzfv/+7777r2bNng0rO6lkRqoJAi2XKBcceP0/M1Taa7Kxb3tovwfhlUVF569bbrFtbQn0KilevsiZMJI+Jy5dzVEq7w4dKPsV04Hbt6hr3VvnhI9PVFROa97QBibd9U2/7pnR7ARhgC1ouWnXvN7LY3r1DoGtb7QrqjAxSnTXIbkSYskDv2LFjx44vq9179uzZ43MetYEDB/bq1atRo0YAgOP4oUOHsrOz7bTCcl2+fDkmJiY+Pt7JyQkA+Hz+xo0b1Wo1g8EAgLFjx44cORIAVqxYsXHjxnfv3iGBrlxOPUygxAgGAP/aJroireDEP9pFtU6qp5KR3YjQLkrDr6nTMxgO9gbwzChgHA6rfj1dO5GfD2iTcYm8ynqVLkmrZ1Nfb2TnNq6BB4OOxGa/tuPZ1bKqHZ32SKwQN7ZvYsu1BQCMT12WaoIBEbUZOnTojz/+CAAEQaSmpi5fvrxt27b3799ns9nz5s27c+fO7t27o6OjL1y4oHtuTExM3bp1nT7vxwkODg4ODtZ86u9fGLMQx3E223irbspHVRDo5FzqYtiGblabL77efPH16Ha1p3Spa1KzhfzBgws+v20BELizni+bMvaN8v17ZoMGutNilI0GAIDxzVvX5A8e5K1eK4uKkgDYbN7IHzyYbo9MDgKIjQ/XX/90jSzO8J3VvaaetLw2XJtWLq2lKuny20sfpT0ijUvbLG/u2Bzj8y1GfCs+clRTmRfc0wielxtHR0c/Pz9N0cvLq2HDhtevX+/QoUPPnj0TEhIGDBgQHBw8YMCAoCDq0L9CocCLz9XJM6t+QFWYJGzsVmT2XMBlvvw83HEwMu7fB5/ocKpYLH/8ntOmcDEWJ7Ct8HvqxHTu8tDUjp0zx41PbRMo2ruX8inle2Xx7XBdyTYjCKUyvd8ATUy17NlzlW/e0OuSCfIw9YFGnQFgW/QWuarYvTAX3p/TqDMAnHxb+MZmvSLUevUq/sABwhnTnR89MMdNT1wuNyIiIiIiIjo6evXq1X379tW7OLpRo0axsbEZGYWJHcLDw5s2bSoSVfrezsqgKvSgu3u7AxoVTgAAIABJREFULDnxFKAwiYRIWiQaw713mQP9PWhyTQ+4m5vd0cPy+/eBwWA1bCg5c1adlcVp347t4wMAimfPRLt2a8Kt5y5dxu8/QHsEg2Fr6/zwvuj3P9QpKdKGDa0n0xDZ0oAo376jWLJmz+UPHmQxcgTa3qYhSURZoUikS9LdBPonh5NERbKhP06LJggCwzBgMi1GfGsx4ttKc9OQpKSk3L1buAkoLS0tNDTUy8urZcuWjx49UqlUGzdu7Nq16+3bt7du3QoA165dGzRoEI7jb9++TU5O7tWrV+3atUNCQpYtW5adnf3TTz95eHgITHtIpziqgkDnFpC9Cf0DGQLTS62CsdmcwEB1Xl5yw8/R11avsV610mLkCF3BUsbFsYsOMePOzlY/LQaAlJQU843LQYLrjJ4rnjzJffJE8eSpzWZTXwJlNDwtaxU1YM7FJxisZVWksp9TC3PcsRIWFqZZd2Ftbd22bdujR49yudw2bdosX758+/bt69evDwgIOHv27LJly2bPnt23b99Ro0bNmjUrJCQkIiLiypUrs2bN6t+/P4ZhPXv2XL9+Pb23U24wwmSXO5QZtZpo8+ul4j49MDmggWth+GC5XC4SiWwr/+VOJBKV+ost2rs3d+kybYtb4ifFy5dpXbtrG12ePi4ucVRKSopz5ceHLMu9VIScxT+JDxzUtbs8j9EN+V9B5HI5QRAczlfsjylfKwZ/zPbG7Dn9eVvKwpaLyS2Lev80KrVyw8P1kYmFk8nrO2ysa6NnVrbsGOcxQ+jFvPtfJAwG9kOvhmvPFuZka13HfnibmjdepXNZjN6+brUcTPTVRpWSqhmWIVHEvlG+ecsfOqTg7+OkxXpFaGWk9TMprEOXc9u1y9m3T3UrSttOiERgaIGuCCni5P3P991Oimru2LyPVz8/J6NuJproPalP7b7pknRPS08hu6SojTiD+YP//GENhosV4lpWtTh45f4aISqVqiDQADCoZQ3vGjbRH7IcLLnt6zsycayll6mvPOP4+4t27dYUWQ0apHUqzLrEC+rBH/4Nq3HjahFAHcO4QT3YTo6S3n01Nk77dprEH6YAAcT2x9vI7dSP0h49Snu0q+teV0F5ciqWG2cLZ2eLsj4PHkITmndBlJuqsIqDpJ6zMKR1zc6NnJi4eYy4cYN6CGdMJ485bdsqXn2Jkym5cBEAqoU6fwbz9LT8Yy8Zso7Xu7f1qlV0e1SE9II07WAXAPAkvUhSBTWh/vPV/5ZG/bz09i//vAurAiOHCFOg2B50bGzswYMHk5KS2rdvP2LECNbnKfWYmJgdO3bs3LmzuBMRZcdy4QLh3DnqvDyioCA1sEhcN+W7OPiaiMlVAHanTkKdNa0mAlcnZzaP+WXrhzo/P/z89hsZl5NcuQDwKPUBpsRC7L4BBKJi6O9B37lzx8fH5/jx4y9evJgwYUJwcLBCoSA/io+P37Vrl96zEOUA43JxR0fd13lW40a0+IPQiyXbsqNHJ22Lv3NL8kAWFZXcoFHDudt/XhE77kA8Q00AYM9zntPhJqKqoV+g58+fP2zYsFevXt29ezcqKur27dtVLMqqqYExmbZ792iKgokTOIFmHCuySjLLd85E78nt3NsPrjf0YNARC1bh/qCMISGaOv73c1rezwYARhUaPETQiP4hjujo6A0bNpCxRVq1arV169apU6eOGzfO09PTqN5VJ3jBPV2exyhevmK6u2unrEaYCEwGs49X395E3/DnKfuvpTpb5fbzc+erFZRqbglSaAWBTm31XgSB+Cr0C7S1tXWKVl7OMWPG7N27d/z48WToVUQlwbC25gS0ptsLRElsOP/y+N3CUONRselrhzenVBA2bLrIf0o9Xn2ju4aoguh/EevTp8/06dOPHz+enJwMABiGHThw4O7duxMmTEhISDCuhwiEHgixmDB6dIUCmVKjzgDw4H1WxKs0u4P7NRZuUI8BP/ze3MFP39kIxFejvwe9cuXKvLy8kJAQPz+/+/fvA0C9evUuX748dOjQAwcOGNXBqo40PFz+4CHu5sYfNFA3BSpCF0Iqzfn+BzJ/I69PH5t1a4wWOTNPQh3QyBbLuV27Okc/kj94gNvbs/1bAIaBUbJ6I6oD+gXa0tLy8OHDO3fuTEr6EnglICDg/fv34eHhsbGx5W7v3bt327ZtS01NbdWq1fTp05lmHkqiguT9tjJ/e2FUcsmZM3b79yGNLpX8zVs02XUlp0/jzk5WS5cYp2knK16LWrYP3mdpLNGiMDzWvY9XPxOP3okwU0qaaxYIBPXqFe7iv379ulQqZTKZ3bt3nzFjRvkaU6vVoaGhvXr12r59e2Zm5vHjx8t3naoBUVCgUWcAkEVESi9cpNEfc0Hx/IV2UfkuzmhNYxjM7dmgdZ3CTaqWNS6/FIcffnFo44N1BKCdKQjDU9bFQH369ElNTa1gY8+ePePz+V27drWxsSGDTlXwgmaNOjOLYlGlf112FdNEFhGZG7pCtGOnOot6gwYBsyiSHATjGDUphpeTcNNIv3UTLR19t3DtCsO/3E6+nVZQFf52VZIjR44EBgYKhcI6deps2rSJND558oRML1t2WCwWmSaxHAQEBLx+/bocJxp1hCE5OVmzUK9GjRopKSmFkWqrJbg7NZ4vp5XZL+EQ7d6Tu6xwyXzuit9cHkcbPB2XcOpUyanTmqKAjojYSoKaxEeqpFoQBoTIF8miooDF4gS2wb4mGOHGjRs3bdq0Y8cOX1/fmJiY8ePHC4XC8ePHV56rFMLDw48ePXrnzp3ynW5UgRaLxZp8M3w+X6VSSaVSjeXNmzcSiQQAcBzX5BMzLGq1WqFQaFItVB4KhUIqLf0by/rfUdWmzep79wCAuXhRrqsLfI1vJnUvJLJlRTY0pe3fzxw3tkwnymRyubyETEVfcHXhnDurvHgRCALv3i3Pw6Ps/2lqtZogiDK1UiLueJFQRI2sG3NlXM3fgvbHTKqSvsmLZQCjvnUDJlbR73jZ74XBYFRGLF/F69i0zoVhDzht2ths34o7OpblxNzc3GXLll29erV58+YA4Orqun79+j/++ENboO/evTtnzpzbt29rH6vV6pkzZ/755584js+ePfunn37q0aOHUqn08vJ68eLFw4cPZ82a9enTp6CgoHXr1jk7O79+/XrKlCldu3b9+++/nzx5ou1DdHQ0n8+3KG/ao7L+8f7880/Hsv2nlICFhQUpwQAgkUhwHOdqzYl5eHhoEthUUjJHhUIhFoutKz+IZUFBAV8nTaceAttAYBt1VhbDygq+XjUUCoUJ3QsAoVBQ3vM5YrGwbB7m5ubyeLyy/t2traFcW+EVCgVBEBV/uqzBOrTNyn/fnXiY+iDQte3w+iPsLL5EhaX3MUsSJ864OY08bmLnPdd3ng3XpiKtGOcxK4H89Rs0x7KoqPxNm61/W1GWE+/fv+/q6kqqM0lISEhISEgJp5D8+++/V69ejY6Ozs7ODggIGD58+MWLF1ks1rt373JzcwcOHLh///62bdsuXLhw1KhR5O6Qx48fN2zY8MiRI5RLzZs3DwBOnjxZFod1KatA9+7du3wNaOPs7Hz27FnyOCEhwdHRUXt8g1v5CxjI1OtGWDryda2U95cPwzDTuhcmk9uju/Til91MvPbtyngug8HAcbyyb4fsQVewlY95H5PESbUsPZcE/FpcKzQ+Zsdi/9QcP8uMOfHu+JRm0yrSinEesxKQfBYNElVCYnE1KcTHx3t4lCfsqlKpJAgiLy/Px8cnISHB0tJS89GZM2c6depE6uGGDRvs7e3JbqVKpdq8eTPL0HnaSvl/V6vVFy9ejI+Pp9gnT55cjsa8vb3z8/Pv3bvn4+Nz8uTJjh07luMiCFPGeunSHJVaeuUKAFguXMCtcgH5/ojZ+9+7wt7QeO+J/bz60+uPLjcTI7WLVWD2kte/n+Tkf5oi06t2GU90dnbW3hENAFKp9K+//ho2bJje+pogsYMHD05OTu7bty+DwZg+fbr2urVPnz6dPXvWxcWFLOI4npaWRrZFqvO2bdt+/vlnANiwYcPYsWUa3yuBUgR6/PjxBw4ccHJysrEp8pZUPoFmMBiLFy/eunXrli1b/P39hwwZUo6LIEwZvIaH3cH9hESCcTjAqGoBg97mvNWoMwD8EbO3o3tHK44JpX0BgPbuHSISbmiKLgIXGp0xCFbz56vTM2S3bgEAp0N74ZzZZTzRz8/v7du3z549a9KkCWm5cuXKwoULR48erV1NqSxMM63ZJv3p06dBgwbNmTPnyZMnY8eOdXNz0wyMODs7h4SE7Nu3DwAIgoiPj3d2ds7NzdVMbMyYMaPcC5F1KUWgT5w4sXHjxtmzZxtqrUWdOnU2b95skEshTBaMR42eXDVIyP9EsSSJkkxNoEc0HKURaB9H35D6Zh+WGq/hYXfkkCImBpgstneTsv/wOzk5zZkzp1+/flu2bPH19X39+vWcOXMo6mltbf3kyZOYmBgPD4/t27eTxrCwsKNHj549e9bGxgbHcdHnoAL5+fnBwcGLFi26ceOGj4/Ppk2bLly4QE4wVhKlCLSdnV2vXr2q7Uo4Y0IoFKoPHxhOTgytAS+ESVHT0pNicTe9zFLOFs4n+p6MzX7NxJh1beoxsKrwHoOx2Wy/8kQ4CQ0NdXFxWbJkycuXL11cXMaPH//jjz9qV6hfv/706dMDAwNr1669ePHiDRs2AMDkyZMjIyPr16/PZrMHDRo0atQoABgyZEjNmjWTk5MPHjw4bdq0Dx8+tG7d+ujRowa5weIoJav3+vXrHz9+vHfvXiPM4BkBk8rqrY3s9p2MwYUDPsLv5lrO+64sZ1WNrN4k2dnZfD7fCPm2K5jV+8jLQ3+//os8ntpses9awXpbMc3HrHygrN40UopAv3jxokWLFhiGubi4aPej37x5U/m+GR4T/eYQRKJ7DW2DfdjxssQdRQL9tVRcoAEgRZxCruKw4ep/kEz0MSsvSKBppJQhjpEjR9auXXv06NFGeNqqLSqtiFQk8kePUGBok+WrsmsjEBWhFIF++/bt5cuXW7ZsaRxvqicMyjpoApjlWryJQCCqGKVMILRu3To3N9c4rlQB4jPFv/4TM/fIw93XP+jGDi4OjMWyWvKLJhoat1tXLopdaW6o1cSnzIJsMYoEjTAkpfSgp0//f3v3HdfE+T8A/Lnswd5D9lQZkSG4F1or1lm1am3rqq122d3aVltHpx2OX7+ttVartda2Tlqte7NERBSQoYJAgECAJGTnfn8chhACBMjm8375x92TS+5zcPl4PPfc51m1cuXKVatWBQYGavZBT58+3ciBWRORVMGiUZrEsrlbLqsbq5pkXy2I03P8i93zy+nJydKMDPKAAcyJKah/l8m2OuX1oi+PF2SV1SOEpg7xfXNymLkjAjaim0RAjOj++GPtR1r5fL6xIrIqd7mCrSeLiG/mtLh21emu3K2rahT7OuseEaysqSE5OWnW5aLGRFNjoo0aLTCSbf/dJc4BhNDxG5XhnuyUCEfzhgRsQzcJGhJxF3AcPfP9VfXq0RztEgEqlY4RMvLCoqa166SXLyOE7F99xeHtt4waJDCBi4XtHqcuqGqGBA0MQncf9ObNmy9cuIAQSuuEaYO0UDXNXRXhTA51G+Ciowhc84YNRHZGCAm+2yI5c8YowRkVjovT/mlav0H0y278UXnC/kw9xwrBy9EWHhqwGeYt2H/06NGBAwfa2dmNHz++sLCwp2/XfQX95ptvvvnmm2PGjOmsqohAIOjpnmyPE0u7ctUL40OLuIJzd2qmxHiunDSwYwc0LhJJzp3XbJFlZFpXRSHxvyeaPv5EWdH60LP45EnXX3b1qIa67Zk91C+9pK1i8swEX6SCu4UGJlOoMAxRyT17MNK8BftramoWLVq0f//+UaNGffnll3Pnzs3Ly+vZR+C6cLnc5uZmnS9ZNalUWl9fb8AP3HqyKOmjE8S/l37JEssURLtAIND9BpXqoc8AzX+CnT/3eu/V1dW9fq/+NI9FfOaMVvwPfQa0HD/e9700NDRIJJK+f07XpFKpkfZSzhMdzHjwT25li1Rh8NOsM52eZgZlmtOsCwKx/P0/colv2aYj+XKFSs83NjY2Ojk5Xb9+Xd3y+++/T5w4Ecfx3NxcDoeD43h6enpycjLxqnpZqVSuXLnS2dnZzc1t/fr1OI5PmjQJIRQQECASiS5evMjhcFxdXRcuXEj8cAoLC8eOHbthw4aYmBjNAA4fPjxq1ChiWSaTkUgkPp/fo2PXfQVtpAlNbM9Lk8I5gc43H/C9HBlTOL4MandF9zHMcf0nTR98hDCEEKKPGMG2qpJ+mlUf1VR1Rp86xPL5ubL8XFsfB5XJejlzHdDp+zPFZ/Jbq4Yeuf7Qx5n57Ci9Ko6avWD/hAkThg8fTiynp6cHBgb2dOqDTm8S4jh+8ODBnTt3lpSUVFVV+fr6hoaGLlu2bPbs2VA7SdPIcPeR4e76b2+3ZDGNEyu9fIXs7s6cPg3Tb7ISS0ZLhocegRH9ldmuHv2tikY932j2gv12dnbEs/hHjx5dtWrV999/39NIdHfoKBSKlJSU+fPn0+n0hQsXfvXVVwsXLqTRaPPmzZs0aZK6fCroHVpcnP0rL7PmP2V12Zk5fRpqPzLF+euvqOEw7FcHZU1N0/oNDS+sFO7YgcugS7r3Rka0uwCyZ+g7a4nOgv27d++WSqU6t8c1Cva/+OKL06ZNCwkJ2b17N65RsEhdsN/b2zs4OFhnwX5nZ2dnZ+ddu3YhhBoaGmbPnr1+/frDhw/3Yl4q3VfQn3zySXZ2dk5OTmxsrGZ7bm7umDFjNm7cuHbt2p7uCfQFrlDIMjJwuYKeEI8Zvz5OZxgTJrj8+H3Lwb+QSkWLjVXcv8d//U3+62/aLV/u+NEHtlehv9fw5mbuiFHEsvjYMfnNPOdtW80bkvV6cqj/5aI69eqC4YF6vtHsBftlMtmkSZMmT5588OBBUq++Hbrfc+zYsTVr1mhlZ4QQh8N57733jh492os9gV5TNTTUL1zEm/tU/cKnqyIGyovumjEY5tSprrt3uf66W9lQ33KodXoR4Y4dwl2/mDEqS6O82G7eqZZDh1XGn+TbViWHuu1bOXzp2JAXJoQdWj06zMtezzeqC/anpaVVVVWdO3eui4L9jY2NmgX7p02bVlVVpbNgf1pa2oULF4gpwzsb50Y4fPiwUqlcvnx5RUXFgwcPHjx4oJ4XW0+6r6Bv374d30l57ISEhI8++qhH+wA1TZK9V+5V88WRvo4LhgWw6D17kluwdZt66DRCSLB5s8uPP6h4PFRdjcxXB1K0e4/mqvTqVbulS8wVjMXpMDxcJWohuencFHQvxNM+xFPfvKzJvAX7c3JycnNzAwMD1S08Hs/V1bXzd2jTXQ8aw7D09PSkpKSOL2VkZBAjUfTfh+UwS6FeoUSR8mnb0yijIz2+mN+zEfL1zy2RnDql2cKaOYO4emWMHeO0+SuyMdN0Z0WH6595TvMpG9a8uc5fb+71XqyoHrQ+e2kuKJBOaetwpI8e5fbbPmTou+tQD9rmQaeh0amrNBAuFtZW8Xv29B0lNERzlRo5UN23IDl/oXnjpj5G2DusecQAwdb/qu2W9HUCY1tC8vNz+20fY9xYhBDziSecPvvM4NkZ9Aed/q398ssvOzrqqCcA1Ud7SipXIoQj1Pb9lCp6Nk7W/uWX5LdvSx91a5I93OQaj4y2/H3IeesWQ0TaM8zUVPejh8Vp/2B0OmvuHEpQkOljsGT0MaPpY0abOwpg3XQn6CeffLKzNzg5OQUEBBgtHhuUEOyqmZ0Tg10DXNndvgtXKJBCgTEYCCGSo6Pbr3tkN27gMhktPr5506dI4x6UGZ8Up8XH924qTwCAPnQn6IMHD5o4DhvmZk/fsSxp96Wyy0V1j8V4rxgfRiJ1+deuUtm4dp1o1y8IIWbqFKfPPyM5OyMKhZaYSLxut+J54c6f1ZvbrVhuxOgBAOYDheENqYovvlcnDHRjO9JaW1QqXKZURfs5fbUgrsu3thHu3iN6NGRNnPYPZmendfON7OvrfSdffDytmcfzmDmT7A/zYwFgmyBBG8zeK/e2/dc6QvmZEf4rJw3cdaH0h7MlCKGR4e6vTxno00nxfi2y9AzN1ZYDf3QcHUFydGQvXCDgco06fqNHWv4+1PLnn9ILF1nzn3Jc8z7J2dncEQFg9SBBG0ZNk2Tbf0XqvuY9V8q9nO2I7IwQuny3DsOwLxfoNbqO5NpuFCBj/DjDhmoM0vQM/suvEMst+3/HxWKX7dvMGxIANgCG2RlGeb1I804gQuhqSbsnxy4V1coVej1EZLesXbFa1oIFfQ/P2CSnT2uuig8fwaFgCwB9BgnaMPw7DMzw7zCdCpWi10+bEhLilZ3p8O479q++4nHyX+bjkw0TojGROlR9wsjdVV4FwCTMO6PK7t27g4ODHRwcZs2aVV9f3/0b2oMEbRiejoxVE8PVq4uG+y0a2W5c8OrHI/X/NLK3t/3LLzm8/Rb1UZEXC8ecNVNz1X7VSnguAxgKjvB/76WtT/94Q/onVyovd/8GDd98882aNWvef//9oqKi7du3f/XVVzt37jRSnB2VlZW9+OKLBw4cyM/PFwgEvakx16Py/tbO2FNdPGxouVRYW84TEVNdNIpk+67c236qKOde73eqEoubPv2M9/Qi3jPPiU+c1HzJ9DOqdEF+9y7/nffql68Q/bYfVyp7uhdrn1FFay8wo4oBHSk59MShKep/Vyuv6PlGs8+osmfPHvWMKgcOHFAv6w9uEhqSrzPT15mJECLKXzmyqPqXRuxM04aN6lF3ktOn3f76k56so0aK2VHCwpw+M89D58C25dblaq5efHhhmM9wfd5o9hlVnn766YULF+I4XlxcfPLkyWHDhul1wBogQRuMTKE6kP4gr7zRgUWdyfGIMlAVG1H7Mp7i48ctM0EDYBo40rdSm9lnVMEwDMOwP/7446mnnvL29s7NzUU9BH3QBvPF8TvbT929VFSbdqNy2a4b3MaeVUTSE9x8A/1NlGu05uoIn5F6vtESZlRBCM2dO5fH473yyitEP0mPQILuE7FMeSi7Yuf50uv3Go7fqNR86dydmo7bc5skN+43NIp6MAGS3QsrNFeZM2f0LlQArNTM0FlLo5fHecYneCW+Fv/6qAH6lqBSz6iibiFmVNGqN9vZjCr37t37+++/9+7de+jQIfXGxIwq1dXV1dXVVVVVt27dImqxas6owufz+Xz+4sWLd+7cScxD6OLismLFips3b8p6OPkZdHH0nlahZy0yZeuoZxxvHdHw07mSn86XEo0fzIiaOsRXn704vPcu2c1NevUqxmSyn3uWxuH0NW5gUDiO/1a490DR7wihKUGpS6KWmTsiW4Nh2PSQGdNDenxpop5RZcuWLUOGDCkqKupiRhU/Pz/NGVX27duXlpamc0aV999//8KFCxwO59tvvz1x4sS1a9c6C8DNzW3VqlUpKSl2dnbffPNNfHw8jUbrbGOdIEH33r83q7p4dWSEx8XC2rf330AIDQ9znzLER52dEUIbDuePifSwZ3Y//SVGodi9+ILdiy/0PWBgDP/cSyOyM7HMoDAWhD2tc0uJQpzPyydhpCi3aBq5Z19U0DvmnVFl+vTp2dnZo0ePFolEw4cP379/f0/j1z2jiq0y7Iwq358u3n2pTLMlleOTlls1Itz9iViPMF/X2d9ebP+OdlWhdyxLivZz6ksAppnqwjTTdljvjCqbMjakV7ddQyV4Jr4b/37H06xKWPXC6da6g9FuMavj33Bj9nUKLJhRxeZZ4hU0juMtLS3G+GSFQiGXy0UikUE+LcKTobmaGOT8+qTg1ycFI4QkEsn5ouoO72j37IYrA/UxEqVSaahj6YJUKsWM/9SJXC4Xi8UKIz8gTtydN+xeKO2/REqVUiwWdzzNdufvUi/f4uX9mr/7+UF9/avINL8a/U8zDMNYHZ4pBX1hiQkawzAmU6/Cbz0lk8kUCoWhPnzsYN+ldZKdF8oQQglBLq9OjlB/skqlcnPUfvh7Csfnn9zWXpFVE8O8XB1Q3zQ3NxvpB6VJpVKZYC9SqZROp1vjFfSU4NTL1W3zJ0zwT6HT6R1Ps2vcq5qrzXID/O5M86sxzWkGdLLEBI0QIpGMMryERCJhGGbAD18+Puy50SHNErmrXbvvPIZhyaFuicGu6gkJH4/1+Whm9JykgHKeKNLHIcCt+0lVumXYY7GEvRh7RyQSCcdxw+4lyj3667HfnnpwSokrR/mOinXnyGSyjj+08X4Tzla03VL2YHv2PQxbOgGAThaaoK0IlULSys4EOpX86TzOoeyKSn7LYF+nKRwfhNBAH4eBPn29cAaWJtQpLNQp7E797Vu8PK6IO8xTx3NuTw9axJPw8upuIoRi3TkLIq2gSCEwO0jQRmTHoGiVTAK26nDJoZ/zfyKWL7ief8xnMllMjnSJdGO6E41uTPd1wz4p5t8lk8ghTqFkDB44At2DBA1AX+E4rs7OCKH8+lv59beI5bcS3lE/WEEhUQa6DjJDfMBqQYI2MJUK/yur4nIhl0mnzUgYkBza16FUwPKJFZ0OOvoy+/ORvqNMMNYC2CRI0Ab2y6WyHx/NdHW+oGb7c4nxQS4Iodpmydb/ik7d4iaFuM5M9Bs70NOsYZobjiMcR7Zy64lFZQ/xiLtRm6Pz1SZZ04Pm+yfu/6tSKZO8h43zH48hyNdALzbyDbEc6uxMUD9t+PmxO6ducRFCGaX17/6eW8wVmCE4S6BQNH34UeUA/0q/gMa33sYlEnMHZBhLopbGuMfqfOl+870Pr6y5Unn5WvW1b3O+Plp6RGuDW7y8vQV7jpYeFsmNPqodWBe4gjYuFY4jhARi+ZW7dZrt14p5YV72ZgrKnAT/+0H4c+sjG6Lf9iubmlx/+J8NTL8HLBJBAAAgAElEQVQS4BC4YcQmvqTBjmb//Y3tpytOEe3rhn1yrvys5pY3a3M1y0po3l386daOfVN+t6f1xxMD6ARX0Ab2zKh2wzZSorwQQrQOsxGy6P30Jr4sK1tzVZL2T93UaXgPS3xZLGeGC5VEfSF65WeJX24a+dnvqX/EecYr8PYPLmr8Z4SjdncXEUJnyttNvwv6OUjQBrZ8XOiycaFDg52Hh7mvfzJmeJg7QohOJc9MaFc4fPygflrcAGNpP5Mmy82tn/cUspUcTfBiekW5RbOobITQMO92w6KDHILVy3KlXOuNjVK+CcID1gK6OAyMSiYtGxsiFHpqVbF5M3VgmJd9zv0GdwfGvOQAF7t+WszMbsli8dFjWo3SzCwUl9CCkP1LqxzefccGejw0jfQdJZKLrlVfxRAW6hQ2L+Ip9Us0Mi3BKzGbm6VuiXaLMUeMwEJBNTujsKUyYwY/FtnNm3VTpnb2qtNnn7IX6S7XaYBdG6EWB0JIoVLsuv3zsdIjCKEnw+fMDXmqRdSi52lWJazaceuH6zXZCKFFg56ZE979jHlqtnSaAZ3gChqYGi021uX/tjWsfEnnq9JLl42XoI3kj7u/H3s0NuPPuwfZZPY49wl6vtfHzmftsI/FCjGdTCdh0OUI2oETovcq+eIP/7yZvPbkewdy7/bbYXO9wpw+3fWXXTpfIjk5mjiYvitpbDe2Mr8+v7MtO8OkMCE7g47gCrqX5ErVZ0dvE8Xqzt2pOXenZuuzibkPGpg08uOxPob6ExoXCuV375I9Pcm+es2PZUUYE1N8KytwkUj83yn+Sy+r29mLnzNfUL1Ean+hQyV1P1EOAPqABN1LZbVCdSlRwsu7W2/1bPvv7t7nE0K76xyUK1XpJbwWqTI+yMXNXkdKl168xJvfWvPMbukSx08+NkTglgVjs1kzZ5C9vZuOHKGy2fYL5lOCg7t/m4WZEDAxk5uhXn0sYLIZgwG2BBJ0L1HIXf1Bejin6k1f1y42EEoU7x3IVaf4rc8mJAa33x7H1dkZISTc+TN9zGjGBH17Nq0LPTmJGhHOYrEoGrfvZEpZFjezRdES687xYHmYMbxuDfMetnHkpxceXiAhbIJ/SpBdsHqa0Y4EMsHxsmM1Im6ES8SkgMfIJPgOgk7BydFLQW7sYWFu14p5Ol9tEGmPb9VyMOOB5gX4viv3tRK0sq5O6y3y/Nu2mqA7EsqFn2VuIqonI4Q+TF6b6DXUvCF1LdotRj1CTtb5mG6JUvJ51qfEcZ2tOHOn/s4bCW+ZKERgheC+RC+RSNjHs2MWjQwaHuY+J8l/7KB2xY84/t3c6XrY0K7+WXqJdqInu2pfgFMCA3sZqxVKKzuuzs4IoW9zvqmX1HexveW7U397f+G+nbd2aB7XhYfnrf24gFHBFXTvOTCpqyaGE8v1QqlKhV8srEUIPTUsYHJ0N3+Sh3s7pD2anxAhNDLCXXsLMtn56838198g1phTHmemTjFU5JavpqVGc1Uga959e9fr8W+aK54+OlZ6ZMetH3W+1CIXuTK66g0D/RkkaMNwtaN/MX9Ii1RBo5ApZKyLLkjCrES/vPLGM7e5xOqLE8I6bsOaN5eWNFSWmUX28aaPGGFjz9d1LdQp9PSD/zRbzlecs9IEjSO8s+wc4x7razfAxPEAKwIJ2jCaxfK88kYmjTwkwFmf7alk0sa5sSvqQ4USRZinPbVDNSUCJTCwX/VsqD0WOPlY6ZFKYaW6Jck72Yzx9IVUKdXZPtQr6ZnBz8LwZ9AFSNB6USjx/50p3nvlHkLo+fGhi0eHaF7O3qpoXP5T6yir5FC3TXPb6gLjODp7m3unqmmACyuV46tV1s7f1QBze9skMkb+ZMTGpSefU7f06Bloi8IgM+I9E4iHuQkfDVuX4JloxpCAtYAErZdfLpUR2Rkh9OPZEnsGdU6Sv/rVXRfK1MvpJbwD6Q/mxLfeM/zs2O0j1x8SyxcKar9cMITa5fg8oObOdP996sHM6gwFrkjwTHSiO5k7ot5bErVMoVLcrMtFCC2IfLpjdm6SNv1WuK+upbZewmNQmCwKK9adMz10Bsy90s9BgtZLQWWT5mpmab1mgr5a3G5I3H1e67wYPIFUnZ0RQuklvMzS+hHhHe4Hgk6wKKyxfuPMHYUB+Nn7rR+xsUnayKbaUToMfMYR/m3O15qX2Aih6zXZdDLt8aBUE4YJLA5czelF6zKGSm7XMLJ9zg31bJ0Rgy/SHg/bILSpqsegRxzpTh2zM0KotqVWKzsTrtdcN35QwKJBgtbLxGhvzdVp8e3uvC8fH6peHhnhPvfRxXWgu3YXMyfAGSF0p7Lpq7SCL47fuVkO1dlBp+D+IYAuDr08FuNtx6D8d6uaQsKmDvEdEtiu1G+Et8O5NSl3KpvsGJRwLwcMQ3IpQghRyaQdy5J+PFuSVcZDCPvkyRg/V1b2vYaXfmmt2vF3VsUX84eMjrTo55iBQTRKG4v5d10YriFOIVovebA8EjwTs2uytNpto3sH9AUkaH2NCHfvovuYSSPHB+ko0B7t57T12YQWqYJFb/1RH8qq0NzgaM5DSNA2L6M6fWPGemJ5rN+41XFvYBrDgDCEvRq3+vei/TUirlghppKpbAp7nP/4oV5JZooXWApI0Kagzs4IIalCqfmSSmXyaIBp4QhXZ2eE0PmKc4leQ0f5ji5pLM6tzXWgO4weMMaR7rgi5gUzBgksEyRoUxsS4HK5qG3UR7i3vRmDASbQLG3WailvLv9PcXLbjS3E6rYbW35LPWBH7aY+bU1Lzf6CfWcrzoz1G/dUxAIfOx+jhAssCSRoQ6rii3++UFovlHraU1ekRDqzW2eGzStv3H2p7MrduglRXs+PDW0QSfdduY8QmpPkv3SMdo8ksDEOdAetlgCHgC+yPtNsOf3gvxmhs1S46mz56YKGAnemR2rwVHta23/eMqVs240txEjq8xXnzlec+2PqnyYIHpgXJGiDEUkVs769qF7lNss3L4wjkzCeQPr8ztbnDM/kc5tb5N8uin95UoSZwgSmhiFsTdKHmn3QSV7aj63zxDyE0P/d3P7f/RNEyy1e3prkD1kUFrFa1lRGZGe1goaCMFa4cUMH5gYJ2mByH7QbM5dewrtXJwz1tM+536DZnlVWT7SbNjpgTkneyb8+vq+YX+zCdA12DEYIJXglZnPbhm0Mch3cIhepszNC6BYvL7M6Qz2Qg9ShVBYMwusP4HdsMCoc12pRqnCEEJNG1mpnULVbgM1zpDsleCUS2RkhtHjw0lh3DrE8K+zJ4T4jhHKR1lsEsraZiIMdQ9TbI4Ri3TmRLgO73WlhQ8Glhxe5ouq+Rg/MBK6g+6SpRX7uTo1CpRoR7s7xb1fHLjHYNcTDHiGUEOSaGOyqnj9ldKSHrzPLDLECS0I8/N0gaWBT2XQyHSHkznKPdedo9mNwPNoyMoVEWR3/xu+Fv/HEPBeGy7zI+XQyXY46nbgHR/h31785W3GGWF3JeWly4ONGOxpgLBje4brPhslkMqFQ6OKiY8ByL1TyxbM1Op2/XzLUkUndeaH0TD431t8xzMtxoI/DpBhvKpnEE0j3Xb3/sL5l8ADHeckBHa+pe4fL5Xp5eRnko7ogFArtupsAt+/4fD6LxaLTDTUfum4ymQzHcRPspXen2UPBw19u/0zMP/tO4nsjfEd2vX0Xv5qc2px1Vz/UbPnziUM0Mk1rM76k4bfCffWSekea47yI+V5sHaeTaU4zoJOpr6DXrl17+/ZtYpnD4XzwwQcmDqAvuI3in86X8kUyZzZt2diQfY/q2xH2Xbn31YK4jXNi2TTK0ZyHN8ubEELn7tR8Pn+Imz391cfgriDoxgD7AR8kf2SQj6rSKKVNqGmp8bP302xRqhTf5HydW3uDWD1TfvrA1D+ZFKZBAgAGYeoEXVlZuXbtWuLiwtgXMoYllSs3Hrmt7qmoaZLQ2hcOJUY3c5skR3PaKthdvluXc79Be8ZuAIws0CFQq6Xj1fEDwQN1dibcrs+HQtUWxaQ3CZVKZWNj4+DBg319fX19fd3c3Ey59z4qqGrWnIc7q6zexb7dH4ypHB+EUINQe/qMekG7lux7DbsulP6TWyVXwkOEwFii3KJnhM5Ur76T+B6VRNXapuNt7X7V4WkVTHoFXVdXR6PRNm3aVF5eHh4evnjxYtcOc1dbkYlR3mKZ8nQ+FyGUHOq2cmI4QijYQ7tPMFrj5uGPZ0t+vlBKLP93q/rL+UOoFNK1Yt6p/GoKiTR1iG+MvxWXpQcWZUnUsqnBT9S21AY4BGo+86IW5BCodVtysOtgEwYIumfSm4SFhYVbt25dvHixt7f3wYMHq6qqvvjiC/WrXC5XJpMhhDAMc3IySp5SKBRSqZTN7s1EU1K58qPDRTkPGonVuACnT2ZE0KlkvkguV+EeGlfTeRVNbxxo7Wd/PzVs3MDWEktShWrqt+man7lmajiO0Kbjd9Utn88ZFBeg77E3NTU5Ojr24lh6RCqVmqAzqqWlhUajUSjGvWJQKBQIIRPspdenWY/0/VdTL+H9de+vC1XnRniNnBk025vl3XEb/U8zDMNMcD+5XzF6gk5LS9u7dy9CaOnSpSkpKep2qVQ6b968PXv2ODi0PgjL4/Hk8tZhQ+pGw1IoFBKJpNfnELdJ8uvVCr5I7symLhru5+XI6GxLiUQiwyn2DIrm4wXVjZKnf2xXl/3FcUG55U3XStueZBkT6fbRtEg942lsbDTS/2SaJBIJg9HpkRqKSCSi0WhUqvaf4YalUChwHDfBXvpymnUBx/FjD44W8gtwhMe6csZ4jGUyjH5PT//TDMMwFguGkBqS0bs4UlNTU1Nbp+25e/euTCaLiopCCFEoFBKJpHktY4IuaZlMplQqe31pE8Jmr3tSrz4ZHMddO3w/gzucu4lhnnmVQs0WFU7SPzyBQGCCyzQcx02wF5lMxmQybWaYXV9Osy4cKvl7f/FeYoafXN4NKok6LWK6wfeixTSnGdDJpDcJpVLppk2bCgoKWlpa9u3bN2jQoH71/y0Jw3Ysa6vw++rkyEG+juHe7f5WGBpixZ3ywNhu8fI051/LqYM5sWycSW8SRkdHz5s37/PPP5dIJDExMW+88YYp995HOI72XC67cZ+PEIryc1oyOphE6vGMy9F+TtfWPVYnkDizacT03kvGBEvkyv1X7yOElo8LnT3Ur5uPAP2Y1iTfZCjHYetMPQ56+vTp06cb/Y8yY/gzs/z708XEcnoJj04hLRoZ1IvPwTDk4dDWpUslk159LAIeYwH6GOYzPIubqV4d4T3KjMEAE4BaHPrKKOVpruY+4PcuQQPQayn+E2kk2uXKSySMNNZvXJR9tOarVcKqY2VHGqWNse6ciQGTyBjU5LJ6kKD1RSG1+3NSq/qjQKL49kRh2o1KhNDqxyNTo6ArGRjF6AFjRg8YQywLhW13mOvEdS+cXk4sX6m8XCEoXx69wgzxAYOCBK2vKRyf8wU16tVJ0e1GjH73KDsjhL75t9CRHjl5CAwIBcaF4/iZ8tN5dTddGC6k9tfLx0qPLhm8lEyCL7h1g9+fvkZHenzzdHxabiWO0OQY71ER7abiPn6jXW2ajFL+5CEBpg0Q2Lg79beLG4u9WF5DvZKIScH3Fu/5t/yfzraX4woyfMGtHPz+dFCp8L+zK64V82gU0swEP/XQt2FhbsPC9BqszaZD9x8wpL0Fe/4oOkAsD/cZ8U7ieypc2SE74+pBeMN9RjDIRn+8CBgbJGgdfrlU9uPZEmL53J2a7xcnDgnUUdv3QmHtsesPlTge6++8amL49lNtT2w/HuNpolhBPyCQCdTZGSF0terKjbobAzvMqDLCZ3SLQnSjNme8/4TFg5eYNkZgFJCgdVBnZ8K/N6s7Juissvp39reWarxWzFs0MujTeZzLd+tYNPLsRH83JlQFAwZTL+FptdS11MZ5xMW5xefw2p5VmRk2K9wZppG1KZCgu6dQ6agLShSxU/v18r30jx8bN6j1wlnz9joAfeTD9tVqCXeOQAgtjlzKvMe8UnkZIfR6/JtdZGe5St6x3CiwfJCgdVg0MujXy22zpUwYrGO+nw6TLANgLDQybd2wT9Zd+4joZV4WvTzIMQgh5MpwfSfxPdRlhf1jpUd33PoBITTUK2l5zApPVqedbzjC5Up5x2mxgBlBgtbh+fGhNAqpoLIZITQ51ntEuHvHbR6L9j6c3TZzynOjg00XH+h/4jzj/552pKaF68p003n3T4WrMrgZvJbaga6DQp3CiMa8ups7bv2PuHNITHXY2ZRaB+8e+PXOHoTQUK+kFbEvujN1nPPA9CBB60Alk5aPC+16myGBLl8/HXco6yGO43FBLvOS2w2qU6rw7afuEpfhy8aGLBkbQoJLbtA3FBLF126AzpeUuPLTjI1ECkYIPTd48aywJxFCN+tuahZXyuRmSJSSjvk9ozqdyM7ENmSM9F6SNU0WasMgQffe8DD34WG6LzR+z3j46+VyYvmn86V2DOpTw2BYNDCWzOoMdXZGCP1ye1dq8BN0Mr3jRCqVgkoyiRRgH4hpXDHk8fI0t7lWfU2pUsBDLpYAqmEZxZ0qgeZqtsZkhsAC4WIx/miyCGtUJ67VammQ1COExvmN02z0YnmvPv/KK2df+vjaR82yZnW7I017whTIzhYCEnRXlCq8oKr5TmWTQtmzYXNa/RkUMvycLZSqqanh+ReqQsOrAoMb330fKRTSq1frn1tSv+iZxjUfqBobzR2gXohBHZo8WV4IIUe606+P//b0wGeeCJnO8RjCbakmXs2pzfmtYK9644kBEzXfuzz6eSPHC/QF/092ii+SffRnHjGTd2Kw60czo9wd9H00a+Jgj/RSvnp1Wrz2MClgIZo3bhKnpRHLol9/JdmxBd//T/2qilvjsnOHmULrgUiXgQsHPr3vUc7dMGIT6VGpaEe649yIeQihT66t03zL1aor/9xLQwi9nfjuSN9Rvz6+79SDUwJZc7xnQox7rEmjB52DBN2pHedKsh51TWSV1f/vbMmHM6L0fO+YCDfXRXYn86pJJCyV4xOn60FEYAlE+35rt/r3Ic1V8YkTuFCIWcNEqPMi5k8NntYgafBme1N0dVAwqe1mL2yUtv5x8EXWZ372/gEOAU+GzzFFoKAnIEF3itso0VzlC2U9entyqFtyqNFnWQR9xJgwQXLmTNu6VKq9BdlqyqqwqWw2tdPJAxdELrz08KLOl/LqcsUK8Yl7/0hV0uHeI0YOGKU1dQswF0jQnXKxazdi347R9rNqapHzRbIBLiwKGc5j68Z+5mnNBE2NiZFebMtiZC8vjGn0abNNw9duwP7UP3Jrb0hV0m+vf635kkAmfPti6/xzVyovN0r5T4RY5bRHtgcSdKeWjQ3RLCKqHhmtHuCMENqxLCnaT68Z6YFlYqSkeJw6Kf7nX4zBYM2aqaypqdNI0ParXzNjbH3ElzQ0yZp87QaoH/JmU9kjfEcihArrC07c/5dojHaLqW2p0XxjTm0OJGgLAQm6U15OzPMfpGSXNeAITwhyZdLICKGrxXWaT4Ev/ykj/ePHzBcjMADqoEHUQYOIZbKPj8fJf0X7f0cyGSM1lTF2jHlj67Udt344VnqUWP581JcDXQdpvvpC7MoIl8iC+jvuLI8pQak/5v1P89XrNdn3m+8V1Be4M93jPRNMFzToABJ0VxhU8siIdo+i3Kls1tqmUSRzYkP5AttBjYpy2rjB3FH0SXr1NXV2Rgi9c+mtozPSNDcgYaQJ/ikT/FOI1aHeSRcenle/GuUa/crZl4jlJO/k5/yhcqnZwPjcnvF21B5p58CEImHAspQ2lmi1CGQCnVsSRvmOXhHzwhCPOITQ1OAn8utvqV/KqE7PbbhhjCCBPuAKumcmRnufyueml7TW53176iASCe4T2r6Wv/4WHz2GMIw1YzpzhqX3z3qytesv2tG6GSmYGvxEavATCKHaltrjZcc0X+JJ6wwbHtAfJOieoVFIXy2Iu1hUy2uWxAW5hHpq1zoAtqflj4P81a8Ty5JTp5yVStbsWeYNSUujtHH37V1nyk8neCZODZk2ZsDYq5VXsmuyiFdfjVut/7C5jnXsAu2CDBkr6AlI0D1GIWPjB8GMVv2I+J9/260eP25pCXp77taM6nSEUHZNVnZN1rYJ/7cm+cPrNdl8CT/KLaqzGng6YRi2fsTGD6+sIVbnhM+LcIg0StBAD5CgAegO3r4Si4VNZyaSi4jsrJbNzfIPC0jwTNxXsPfF0ysQQjNDZz0z6NmuSyApceWlhxerRVUhTqEHn/i7WlTlwnB1oDlwudwu3gWMChI0AN1gjBsrOX26bXXCeDMG01HHB7tpZDpC6Gjp4YN3W6eaPVTyN4PCnB+5oLMPUeGqzzI3qRP95MDHV3JeMk68oAcgQQPQDfazz+AKhfTCBYQQY/x49tMLzR1RO3QyPcV/4unyU+qWYT7DEUK3eLc0Nyvm30UISZXSf+6llTc/CHQMmhz4OJ1MJ169XZ+veRl+4v6/cyLmwbwqZgcJGoDuYJjdsqV2y5aaO45OvchZ5efgf5uX78xwnhE6y5XhihDSmiWWhJGUuPLzzE/VNw/z6m6+m/hehbCCTqY3SBq0PrNBXA8J2uwgQQNg9agk6szQWTND2926nBQ4+WrVFfXqBP+U0sYSdXZGCGVxM9+9/HYxvxghxPHgaH1mgGOgESMG+oEEDYBtivOI+3TUF+fKz+AIHzNgbIx7bE5tTvtNcCI7I4Rya3OHeY+4Vt2a0NeP2KhzalpgYpCgAbBZg10HD3YdrF4NezTb9yPtBkfjSHV4+rEmaaMTwxnKjVoIeNQbgP7Cnmb/2agv4jziEEJxnvGDNHI3QsiL7U3CSM4MF8jOlgOuoAHoRwa5Dn4/6cOChjskjExCpPcuv61+aXbYbDMGBnSCBA1AP8IVcZ8/1TocJcY9duv4/7vffI9Gpsd7xNPI2kUZlbjycuWlu9y7HMRJ9Bpq8mABJGgA+pO9BXvUy3l1N/+5d/zF2FU6t8Rx/IvMT69VX0MIHXt45PGgKZ1tCYwH+qAB6EcuPryguVrX0mmlutv1+UR2Jvx77586MZS1MzULvYJWKBTG+FilUqlSqYz04Vo7MsFecBy3mWNRqVQm2JFSqTTBD82ST7PRvmMuVrblaFeGa2efUCfSTse1wlpnqnPXn0+hWGhKsVKW+NPEcVwulxvjkxUKhUqlMtKHa1IqlSbYi/F+UJpMcyxERsMw444fUCgUOI6bYC8We5o9GTxXnaCjXKNnBT0pl8uVuLJcUI5hmDPN+WTFv3XiunCniCiXwVrv9WF4d707DMMgQRsWhuMWVpvLmGQymVAodHFxMfaOhEKhnV03JdL7jsvlenlpl2Y3ONMcC5/PZ7FYdDrdqHuRyWQ4jptgL5Z8msmUMmIUxyCXgWQShS9p+Pr65pt1uVqbjfefEOMeq57/e93w9cT4PGBK8N8dAP0LjUyLdW97sHvPnd0dszNC6Gz5maVRyw9NP1pcURzhHwGDo80CbhIC0K81SZs6e0kkF5ExshPNCbKzuUCCBqBfY1KYOttj3TmebJg5yMwgQQPQr82P1FHeepjP8BdiV8KFs9lBHzQA/ZdSpRhgP2B/6h+5tTdIGGmIxxBGJxfUwCwgQQPQL5Q0Ft+pv+3GdE/yTiZj5JP3T2zP3YoQSvQaujRq+QjfkeYOEOgACRoA23e45O+f83cSywleiU+GziGyM0Ioi5upVCnWDV9vvuhApyBBA2Dj5Cq5OjsjhLK5WUxyu36MnNqcZlmzA83B5KGBbkCCBkBfOMIPFf+VV3cTYViEc8TciKfIGNncQXWvUdqo1aLAtR/vZlFYpgoH9AAkaAD0deLeP7/c3kUs59RcJ2OUuRHzzBuSPtyYblotEwMmXau6ql6dFfYkhQSpwBLBbwUAfV2vua65WthQYK5IegRD2KejPn/v0jvE6vzIhQmeifum/H7y/r/1koYYt5hkn2HmjRB0BhI0APrSusw0dtElAxrsGnVo2pFqEdeF4cyishFC9jT7J8Pnmjsu0A14UAUAfU3wT9FcHek7ylyR9AKZRBlgP4DIzsBawBU0APpK9Bq6bvj6Mw9O4QgfNWDMMG/oGQDGBQkagB6I84izvaqbSpVib8GvfxX/GecRF+sxZEboTHjI20JAggagv/u96Pe/iv9ECOXU5uTU5rCp7EkBj5k7KIAQ9EEDAEobSzRXs7lZ5ooEaIEEDUC/174/g0qimikOoA0SNAD93XCfEQi1TX03MWCSGYMBmqAPGoDunX7w38XKi7m1Nx4LnLx48BIbG6yW4j+RTWFfqrxIIVEmBjwW7RZt7ohAK0jQAHQjpzZny43viOWT908oVcpX4l4zb0gGN8xn+DCf4eaOAmiDLg4AupHFzdBcPV1+CtfoEADAeCBBA9ANOpmh1QLDhIFpQIIGoBsT/CdoruqcxA8AY4A+aAC64Wfvv2X8trSy40K5MNFr6LgB480dEegvIEED0L1Ah6BVnJfNHQXod6CLAwAALBRcQQNg48QKcWFDAY1Mi3QZaBVzdAE1SNAA2LIHzQ9ePrsSIRwhjOMx5O3Ed+2oduYOCugLujgAsGW/FfyKECLKbeTW3vjz7kHzxgN6BBI0ALbsWvU1zdWHggpzRQJ6ARI0ADpUi6oOlfydVna8Sdpk7lj6ROsBbj97P3NFAnoB+qAB0HazLvfDK2uI5R/yvv9h4g5vto95Q+q1hQOfvlZ1lVjmeAyZHT7HvPGAHoEEDYC2oyWHNVf/Lv7LegdB+9sH/PHEX4X1raM4SBj80WxNIEED0I7o11+z7NtNKcKX8M0VjD74Ev7xsmM8cd1gt6gJ/ikdB9IxyAyOxxCzxAb6CBI0AG3khYWN774/bJHftWRndWOkS6QZQ+qaQCZ49sTTxPK5irN3+XdfstqLfWyzjY4AAAspSURBVNAR/L0DQBvZ9RyE0MzD1VG3m4mWMQ7xM0JnmTWormS0H6Tx3/0T12uylbjSXPEAwzL6FfRbb7312muv+fr6EqulpaXbtm2rqalJSkpatWoVhQKX8MCCkL08EUL2AsWq/7svsKdQFHhw1l4SyXLPUpGiRavl42trh3jEvZ34Ltu2pn3pn4x4BZ2Xl7dly5aioiJ1i0ql2rBhQ2pq6vbt2+vr6w8ehDHzwLIwxoxhpKQQy/YChfszS0nOzl2/xfRUuIonrlOqFAghjruOzuUbtTmHSv4yeVzA8Ix4aVBWVkan0xmMtmLn+fn5LBYrJSUFITRv3rxt27bNnz/feAEA0GMUisvOHeKjx5Tl5bT4OPqoUeYOSNv1muyPr60lllcMfjE1bOq7Q9//LHOT1mYVzeUmDw0YnhET9IwZMxBCGRlt0wVVV1cHBgYSy/7+/lwuF8dxDGudnEIoFCqVSoQQhmF0Ot0YIcnlcqVSKZVKjfHhWjsywV5UKpXNHItSqZTJZMbei1wux/HuZ6sip04hRkL07sCNd5pJFGJ1dkYI/XD7+yi3qHjXhIOP//1lzueZNW3fNV+2n6EC0P80wzCMRqMZZKeAYNLONZFIxGQyiWUWi6VUKiUSibqFx+OJxWKEEIlEcnd3N0YASqVSLpeLRCJjfLgmmUymTyLoI6VSaTPHQuxFLpcbdS/EFYAJ9mKk06y0uUSr5Rb3lr2nA0Jo2oDp6gQd7Rw9wTPFUAHof5qRSCRI0IZlyASdlpa2d+9ehNDSpUtTHnXkaWKz2UQKRgiJxWIymazZAaK+uDYemUwmFApdXFyMvSOhUGhnZ/SaYTKZzGaOhc/ns1gsI/3lpEb8N2CCvRjpNFMxg7RaAt0DiR25uLj84fNXUUMhncyIcI5Q/2Had6Y5zYBOhkzQqampqampXWzg5eWVlpZGLD98+NDDw8OApxEANs+N6Z4aPDWt7DixOtQjabBrlPpVBpkR684xU2jAKEzaxREdHS0QCDIzMzkczuHDh8eOHWvKvQNgLZqkjXvu7D714L+hXknTQ2dGu0WrX1oR82Kk88DipruRzgNjHGLhEse2mTRBk0ikNWvWbN26dcuWLYmJiXPmQN0WAHTYemNLJjcDIZTJzcjkZvwvZYePnQ9CSKlSfHfj2/MV5xBCD9wf+IT7Otg7mDlWYExGT9A///yz5mpoaOh3331n7J0CYL0EMgGRndWyazKn2c1ACB2/d5zIzgihm3W5+7F9a9w/NEOIwFQs9xEpAPobkVx08v4Jrqhaq51CohILd3j5mu0ZtekmigyYCSRoACxCi6JlU8aGW7y8Rw04MU8VQijZexixIJALzREaMBtI0ABYhPSqaxrZGSGExXnGu9BdZoU96cJoHeU2wG5APu+W5rsUKgXFgkuFgD6CXy0AFqFZ1qzV8hLnFTemm2ZLvGfCifv/qlfj3OIhO9s2KDcKgEWIcY/RXI1152hlZ4RQknfy/MiFxHKCV+IzEc+ZJjZgLvDfLwAWIdgx5NW41d/lfIMQSvBMXBy1ROdm8yMXzI2YJ1PKmBSmUAhd0jYOEjQAlmKCf8oE/5Ruu5XJGJlJYZosKmBG0MUBgGWBbmWgBgkaAAAsFCRoAACwUJCgAbBKuFCI8/nmjgIYFyRoAKwMrlDwX3+jKmJgc/Lw+meeU9XxzB0RMBZI0ABYGdFPO1sO/EEsS86cadqoPSEhsBmQoAGwMrLsbM3VloMHzRUJMDZI0ABYGZKHh+YqY/x4c0UCjA0SNABWxm7F85qr7MXPmSkQYHQwJB4AK0MJCPDOvdFy5IhUIHB8YiolNNTcEQFjgQQNgPUhubvZLVuKhEKK8SdcB2YEXRwAAGChIEEDAICFggQNAAAWChI0AABYKEjQAABgoSBBAwCAhYIEDQAAFgoSNAAAWChI0AAAYKEgQQMAgIWCBA0AABYKEjQAAFgoSNAAAGChIEEDAICF6l8JWiqV1tXVmWBHbDbbBHupqKgwwV5Mcyx1dXVisdjYe6HRaHQ63dh7kUqlPJ4pJnK1pdMM6NS/EjSO4yqVygQ7wjDMBHuxpWPBcdwEezENHMeVSqUJdmRLpxnQqX8laAAAsCL9K0GTyWQGg2HuKAzGNH/hmgaDwSCTyeaOwjDgNAOGgtnSn5YAAGBL+tcVNAAAWBFI0AAAYKH6RYLOzMxcuXLl3Llz16xZ8/DhQ6KxtLR09erVCxYs+O677xQKhXkj7IW33nqrsrJSvWq9h2O9kWuxmd+ITX5frJTtJ+jGxsavv/56yZIlu3fvHjRo0Oeff44QUqlUGzZsSE1N3b59e319/cGDB80dZg/k5eVt2bKlqKhI3WK9h2O9kWuypd+I7X1frJrtJ+iioqKgoKCEhAQmk/nUU0+Vl5eLRKL8/HwWi5WSkuLs7Dxv3ryLFy+aO8weKCsro9PpmuMErPdwrDdyTbb0G7G974tVo5g7AKOLjY2NjIwklgsLCz09PdlsdnV1dWBgINHo7+/P5XJxHDfNsP++mzFjBkIoIyND3WK9h2O9kWuypd+I7X1frJrtX0EzGAxHR0eEUGZm5ubNm5cvX44QEolETCaT2IDFYimVSolEYs4o+8Z6D8d6I++a9R5Xf/i+WBHbvIJOS0vbu3cvQmjp0qUpKSkCgWDbtm11dXXvv/9+aGgoQojNZqsrP4jFYgt/skDrcDpuYF2Ho8l6I++aVR+XtX9fbIltJujU1NTU1FRiWaFQrF27Ni4u7t1331X/Uebl5ZWWlkYsP3z40MPDw5L/XtM8HJ2s63A0WW/kXbPe47KB74stsf0ujvT0dJVKNWnSpLq6utra2traWhzHo6OjBQJBZmamTCY7fPjw2LFjzR1mn1jv4Vhv5F2z3uPqD98XK2L7j3rv3r37r7/+0mzZt2+fvb19SUnJ1q1b6+vrExMTV61aRaFY2R8TS5YsWb9+va+vL7FqvYdjvZFrsY3fiK1+X6yU7SdoAACwUrbfxQEAAFYKEjQAAFgoSNAAAGChIEEDAICFggQNAAAWChI0AABYKEjQoJfmzJmDPUKj0eLj49etW6euFFxTU4Nh2ObNm/u+o4EDB16/fr3vnwOA1YHR5qD3EhMTt27dihASCATZ2dmffvrppUuXTp8+jWEYi8V68803ExIS+vL5crn8u+++KywsNFC8AFgZeFAF9NKcOXMaGxtPnTqlbsnNzY2Pj9+/f//cuXP7/vk//vjjq6++SlRNy87Ojo+P7/tnAmBdoIsDGAyHw0lNTd25cyexam9vT1TYcXZ2Pnr06OTJk52cnBISEu7cubN58+bg4GBHR8clS5aoVCqdnzZr1qzs7OzTp0+b7gAAsDCQoIEhcTickpKSju1vvPHG66+/fvLkSYlEkpCQkJube+LEibVr1+7atevkyZM6P8rNzW3w4MHq4vEA9EOQoIEheXp6VlVVdWxfuXLlpEmTkpKSnn32WQzDduzYER4evnr1ak9Pz/LyctPHCYBVgAQNDKm2tlZdzk1TUFAQseDi4uLt7U2Ue8cwzNnZmWg/fvy4/SPnz583VbwAWDQYxQEMKTc3NyQkpGO7Zn13nbXeJ0yYkJ+fTyx7enoaKTwArAskaGAwubm5x48fP3DgQC/ey2QyAwICDB4SAFYNEjTovaamJmIqa6FQmJ2dvWnTpvHjx8+ePdvccQFgIyBBg97LyspKTk5GCFGp1KioqNdee+3DDz+E2eoAMBR4UAUAACwUjOIAAAALBQkaAAAsFCRoAACwUJCgAQDAQkGCBgAACwUJGgAALBQkaAAAsFCQoAEAwEJBggYAAAv1/77vvNXi1GF2AAAAAElFTkSuQmCC\n"
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%%R\n",
    "set.seed(100)\n",
    "#pdf('tSNE-All-Batch.pdf')\n",
    "tsNE_out_2PCs <- plot_tSNE(cdFiltTopSel, 14, groups=CellClust, perplexity=30, \n",
    "                              cellName=colnames(cdFiltTopSel[,4:dim(cdFiltTopSel)[2]]),ret.val=TRUE, text.label=F)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Clustering purified cells with mixed population\n",
    "We now cluster the two batches of OE19 single cells and the HET1A single cells from purified cells. This is to confirm that the clusters that we have identified from the mix population groups with their corresponding cell-type from the purified cell types.\n",
    "\n",
    "We will use the same `mergedPeaksFilt.bed` that we have prepared from the mixture of the cell-population."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### HET1A cells peak accessiblity\n",
    "We first identify the peak accessiblity for HET1A cells."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [],
   "source": [
    "%%R\n",
    "mergedPeakFile = '/home/mqbsxsm2/Rna-seq_Data-Analysis/syed_scATAC_ReDoing/All_Filtered_BAM/MergedFilt_ReDo.bed'\n",
    "HET1ApeakFileFolder = '/home/mqbsxsm2/Rna-seq_Data-Analysis/syed_scATAC_ReDoing/HET1A_Batch1_Filtered_MACS2_files/'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "metadata": {},
   "outputs": [],
   "source": [
    "%%R\n",
    "peakAccessibility(mergedPeakFile,HET1ApeakFileFolder, '.narrowPeak', HET1ApeakFileFolder, 'HET1A_ReRun_peaks.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [],
   "source": [
    "%%R\n",
    "HET1A_peakFile <- '/home/mqbsxsm2/Rna-seq_Data-Analysis/syed_scATAC_ReDoing/HET1A_Batch1_Filtered_MACS2_files/HET1A_ReRun_peaks.csv'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [],
   "source": [
    "%%R\n",
    "cdHET1A <- read.csv(HET1A_peakFile, sep=' ')\n",
    "HET1Acellnames <- c(colnames(cdHET1A[,1:3]), paste0('HET1A-B1_C',gsub('X','',gsub('_.*','',colnames(cdHET1A[,4:dim(cdHET1A)[2]])))))\n",
    "colnames(cdHET1A) <- HET1Acellnames"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### OE19-B1 cells peak accessiblity\n",
    "We first identify the peak accessiblity for OE19-B1 cells."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [],
   "source": [
    "%%R\n",
    "mergedPeakFile = '/home/mqbsxsm2/Rna-seq_Data-Analysis/syed_scATAC_ReDoing/All_Filtered_BAM/MergedFilt_ReDo.bed'\n",
    "OE19_B1_peakFileFolder = '/home/mqbsxsm2/Rna-seq_Data-Analysis/syed_scATAC_ReDoing/OE19_Batch1_Filtered_MACS2_files/'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "metadata": {},
   "outputs": [],
   "source": [
    "%%R\n",
    "peakAccessibility(mergedPeakFile,OE19_B1_peakFileFolder, '.narrowPeak', OE19_B1_peakFileFolder, 'OE19_B1_ReRun_peaks.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [],
   "source": [
    "%%R\n",
    "OE19_B1_peakFile <- '/home/mqbsxsm2/Rna-seq_Data-Analysis/syed_scATAC_ReDoing/OE19_Batch1_Filtered_MACS2_files/OE19_B1_ReRun_peaks.csv'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [],
   "source": [
    "%%R\n",
    "cdOE19_B1 <- read.csv(OE19_B1_peakFile, sep=' ')\n",
    "OE19_B1cellnames <- c(colnames(cdOE19_B1[,1:3]), paste0('OE19-B1_C',gsub('X','',gsub('_.*','',colnames(cdOE19_B1[,4:dim(cdOE19_B1)[2]])))))\n",
    "colnames(cdOE19_B1) <- OE19_B1cellnames"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### OE19-B2 cells peak accessiblity\n",
    "We first identify the peak accessiblity for OE19-B2 cells."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {},
   "outputs": [],
   "source": [
    "%%R\n",
    "mergedPeakFile = '/home/mqbsxsm2/Rna-seq_Data-Analysis/syed_scATAC_ReDoing/All_Filtered_BAM/MergedFilt_ReDo.bed'\n",
    "OE19_B2_peakFileFolder = '/home/mqbsxsm2/Rna-seq_Data-Analysis/syed_scATAC_ReDoing/OE19_Batch2_Filtered_MACS2_files/'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "metadata": {},
   "outputs": [],
   "source": [
    "%%R\n",
    "peakAccessibility(mergedPeakFile,OE19_B2_peakFileFolder, '.narrowPeak', OE19_B2_peakFileFolder, 'OE19_B2_ReRun_peaks.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {},
   "outputs": [],
   "source": [
    "%%R\n",
    "OE19_B2_peakFile <- '/home/mqbsxsm2/Rna-seq_Data-Analysis/syed_scATAC_ReDoing/OE19_Batch2_Filtered_MACS2_files/OE19_B2_ReRun_peaks.csv'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {},
   "outputs": [],
   "source": [
    "%%R\n",
    "cdOE19_B2 <- read.csv(OE19_B2_peakFile, sep=' ')\n",
    "OE19_B2cellnames <- c(colnames(cdOE19_B2[,1:3]), paste0('OE19-B2_C',gsub('X','',gsub('_.*','',colnames(cdOE19_B2[,4:dim(cdOE19_B2)[2]])))))\n",
    "colnames(cdOE19_B2) <- OE19_B2cellnames"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {},
   "outputs": [],
   "source": [
    "%%R\n",
    "cdAll <- cbind(cdFiltTopSel,cdHET1A[,4:dim(cdHET1A)[2]],cdOE19_B1[,4:dim(cdOE19_B1)[2]],cdOE19_B2[,4:dim(cdOE19_B2)[2]])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[1] 17013   408\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%%R\n",
    "cdAllSel <- cdAll[rowSums(cdAll[,4:87])>2,]\n",
    "cdAllSel <- cdAllSel[rowSums(cdAllSel[,88:176])>2,]\n",
    "cdAllSel <- cdAllSel[rowSums(cdAllSel[,177:252])>2,]\n",
    "cdAllSel <- cdAllSel[rowSums(cdAllSel[,253:322])>2,]\n",
    "cdAllSel <- cdAllSel[rowSums(cdAllSel[,323:408])>2,]\n",
    "print(dim(cdAllSel))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {},
   "outputs": [],
   "source": [
    "%%R\n",
    "cellInfoWithClust <- data.frame(CellName = colnames(cdFiltTopSel[,4:dim(cdFiltTopSel)[2]]),\n",
    "                               clusterInfo = CellClust)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "metadata": {},
   "outputs": [],
   "source": [
    "%%R\n",
    "cellTypes <- c(CellClust, \n",
    "               gsub('_.*','',colnames(cdHET1A[,4:dim(cdHET1A)[2]])),\n",
    "               gsub('_.*','',colnames(cdOE19_B1[,4:dim(cdOE19_B1)[2]])),\n",
    "               gsub('_.*','',colnames(cdOE19_B2[,4:dim(cdOE19_B2)[2]])))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "%%R\n",
    "cellTypes <- gsub('HET1A-B1','HET1A', cellTypes)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n"
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%%R\n",
    "MDSClust <- plotMDSClust(cdAll, 4, Clusters=cellTypes, ret.val=TRUE, text.label=FALSE, title=\"\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "png \n",
       "  2 \n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%%R\n",
    "pdf('MDS_withAll.pdf')\n",
    "MDSClust <- plotMDSClust(cdAll, 4, Clusters=cellTypes, ret.val=TRUE, text.label=FALSE, title=\"\")\n",
    "dev.off()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 325,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n"
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%%R\n",
    "set.seed(100)\n",
    "#pdf('tSNE-All-Batch.pdf')\n",
    "tsNE_out_2PCs <- plot_tSNE(cdAll, 14, groups=cellTypes, perplexity=28, \n",
    "                              cellName=colnames(cdAll[,4:dim(cdAll)[2]]),ret.val=TRUE, text.label=F)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 298,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "png \n",
       "  2 \n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%%R\n",
    "set.seed(100)\n",
    "pdf('tSNE-Updated_All.pdf')\n",
    "tsNE_out_2PCs <- plot_tSNE(cdAll, 14, groups=cellTypes, perplexity=28, \n",
    "                              cellName=colnames(cdAll[,4:dim(cdAll)[2]]),ret.val=TRUE, text.label=F)\n",
    "dev.off()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We could clearly see from the MDS and the t-SNE plot that Cluster-2 groups with OE19 cells, Cluster-3 groups with HET1A and Cluster-1 remains independent which we believe would be the OE33 cells. There are some mixture of cells in each of the groups, but we would still take the clusters that k-medoid gave."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Generate Merged data for the clusters\n",
    "Now we will have the merged data for each of the clusters"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "For Cluster-1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 101,
   "metadata": {},
   "outputs": [],
   "source": [
    "BAMFolderToMerge = '/home/mqbsxsm2/Rna-seq_Data-Analysis/syed_scATAC_ReDoing/Group1_BAM_files/'\n",
    "getAggregatedPeak(BAMFolderToMerge, BAMFilePattern)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "For Cluster-2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 102,
   "metadata": {},
   "outputs": [],
   "source": [
    "BAMFolderToMerge = '/home/mqbsxsm2/Rna-seq_Data-Analysis/syed_scATAC_ReDoing/Group2_BAM_files/'\n",
    "getAggregatedPeak(BAMFolderToMerge, BAMFilePattern)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "For Cluster-3"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 103,
   "metadata": {},
   "outputs": [],
   "source": [
    "BAMFolderToMerge = '/home/mqbsxsm2/Rna-seq_Data-Analysis/syed_scATAC_ReDoing/Group3_BAM_files/'\n",
    "getAggregatedPeak(BAMFolderToMerge, BAMFilePattern)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "For HET1A-B1... for HET1A-B1, we would need to modify the macs2 calls to run the single-end data."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 108,
   "metadata": {},
   "outputs": [],
   "source": [
    "BAMFilePattern = '_nodup_noexclu_sorted.bam'\n",
    "BAMFolderToMerge = '/home/mqbsxsm2/Rna-seq_Data-Analysis/syed_scATAC_ReDoing/HET1A_Filtered_BAM/'\n",
    "getAggregatedPeakSE(BAMFolderToMerge, BAMFilePattern)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "For OE19-B1. Similarly for OE19-B1, we would need to modify the macs2 calls to run the single-end data."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 109,
   "metadata": {},
   "outputs": [],
   "source": [
    "BAMFilePattern = '_nodup_noexclu_sorted.bam'\n",
    "BAMFolderToMerge = '/home/mqbsxsm2/Rna-seq_Data-Analysis/syed_scATAC_ReDoing/OE19_Batch1_Filtered_BAM/'\n",
    "getAggregatedPeakSE(BAMFolderToMerge, BAMFilePattern)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "For OE19-B2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 111,
   "metadata": {},
   "outputs": [],
   "source": [
    "BAMFilePattern = '_nodup_sorted.bam'\n",
    "BAMFolderToMerge = '/home/mqbsxsm2/Rna-seq_Data-Analysis/syed_scATAC_ReDoing/OE19_Batch2_Filtered_BAM/'\n",
    "getAggregatedPeak(BAMFolderToMerge, BAMFilePattern)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 104,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'_nodup_sorted_cleaned.bam'"
      ]
     },
     "execution_count": 104,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "BAMFilePattern"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Correlation wth Bulk dataset\n",
    "\n",
    "Now in order to find out which of the clusters correspond to which cell type we would calculate the correlation of each cluster with Bulk dataset of HET1A, OE19 and OE33. We have already writen the filtered peaks in the `mergedPeaksFilt.bed` file. We have also merged the BAM files for all cells in each individual cluster and call peaks on those merged files. All the peaks are then saved in the folder `Group_Peak_Files_with_Sig_Corr_Cluster`. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Calculating peak accessibility for each clusters."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 128,
   "metadata": {},
   "outputs": [],
   "source": [
    "%%R\n",
    "mergedPeakFile = '/home/mqbsxsm2/Rna-seq_Data-Analysis/syed_scATAC_ReDoing/All_Filtered_BAM/MergedFilt_ReDo.bed'\n",
    "GroupPeakFolder='/home/mqbsxsm2/Rna-seq_Data-Analysis/syed_scATAC_ReDoing/AllCluster_Bulk_Files/'\n",
    "peakAccessibility(mergedPeakFile,GroupPeakFolder, '.narrowPeak', GroupPeakFolder, 'All_sc_and_Bulk_Peaks.csv')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Reading peak files for clusters."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 275,
   "metadata": {},
   "outputs": [],
   "source": [
    "%%R\n",
    "cdAll <- read.csv('/home/mqbsxsm2/Rna-seq_Data-Analysis/syed_scATAC_ReDoing/AllCluster_Bulk_Files/All_sc_and_Bulk_Peaks.csv',\n",
    "                           header=TRUE, sep=' ')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We now calculate the correlation agains each other."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 276,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       " [1] \"seqnames\"                \"start\"                  \n",
       " [3] \"end\"                     \"Cluster1_Filtered_peaks\"\n",
       " [5] \"Cluster2_Filtered_peaks\" \"Cluster3_Filtered_peaks\"\n",
       " [7] \"HET1A.B1_Filtered_peaks\" \"HET1A.Bulk_peaks\"       \n",
       " [9] \"OE19.B1_Filtered_peaks\"  \"OE19.B2_Filtered_peaks\" \n",
       "[11] \"OE19.Bulk_peaks\"         \"OE33.Bulk_peaks\"        \n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%%R\n",
    "colnames(cdAll)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 278,
   "metadata": {},
   "outputs": [],
   "source": [
    "%%R\n",
    "colnames(cdAll) <- c(colnames(cdAll[,1:3]), gsub('_.*','',colnames(cdAll[,4:dim(cdAll)[2]])))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 280,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       " [1] \"seqnames\"   \"start\"      \"end\"        \"Cluster1\"   \"Cluster2\"  \n",
       " [6] \"Cluster3\"   \"HET1A\"      \"HET1A.Bulk\" \"OE19.B1\"    \"OE19.B2\"   \n",
       "[11] \"OE19.Bulk\"  \"OE33.Bulk\" \n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%%R\n",
    "colnames(cdAll) <- c(colnames(cdAll[,1:6]), 'HET1A', colnames(cdAll[,8:12]))\n",
    "colnames(cdAll)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 281,
   "metadata": {},
   "outputs": [],
   "source": [
    "%%R\n",
    "M <- cor(cdAll[,4:dim(cdAll)[2]], method=\"pearson\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 282,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAeAAAAHgCAIAAADytinCAAAgAElEQVR4nOydd1gUxxvHv7t3HBxX6L2jYsGODUVssSSa2Cv2rmhMLLEllmhiTOwticZYfokaC3aT2BUjYKNZQOkdpMPBAXe3vz/uBMRrKncizueZh2dvZm6+szvLe7PvTqEYhgGBQCAQ6h70u64AgUAgEJRDDDSBQCDUUYiBJhAIhDoKMdAEAoFQRyEGmkAgEOooxEATCARCHYUYaAKBQKijEANNIBAIdRRioAkEAqGOQgw0gUAg1FGIgSYQCIQ6CjHQBAKBUEchBppAIBDqKMRAEwgEQh2FGGgCgUCooxADTSAQCHUUYqAJBAKhjkIMNIFAINRRiIEmEAiEOgox0AQCgVBHIQaaQCAQ6ijEQBMIBEIdhRhoAoFAqKMQA00gEAh1FGKgCQQCoY5CDDSBQCDUUYiBJhAIhDoKMdAEAoFQRyEGmkAgEOooxEATCARCHYUYaAKBQKijEANNIBAIdRRioAkEAqGOQgw0gUAg1FGIgSYQCIQ6CjHQBAKBUEchBppAIBDqKMRAEwgEQh2FGGgCgUCooxADTSAQCHUUYqAJBAKhjkIMNIHwwRMU9NLH0FD06PGOqkJ4CWKgCYQPnhEjcO4cAKSlYfJkdOuGbt3edZ0IAMB+1xUgEAjvmosX0b8/jh9HQACGDUNUFOzt33WdCABAMQzzrutAIBDeNamp6NcPfftiw4Z3XRVCFcRAEwgfMIMHVx3n5eHGDfTtCy4XAE6efFeVIlRCDDSB8AFz/rzKpP799VgPgnKIgSYQCIQ6CnlJSCB8wLRpozIpNFSP9SAoh/SgP1SCguDtXfUxNBTz5+PatXdXIcK7IDxcZVKrVnqsR21TX25vMg76Q4UMfSUAaNVKZXivqS+3N3FxfKiQoa+ESqp3NiupMb3w/aK+3N7ExfEBQ4a+EuSEhCgOGAYpKdi5E/7+GDbsndbprakXtzcx0B8eZOgrQT15eejZ8319SVi/bm9ioD88yNBXgnoiItC1KwoK3nU93oj6dXsTA00gfPBU90FLJAgPx5w52LTp3VWIoIAY6A8PMvSVUINKH7QcU1M0bvyOqvLW1K/bmxjoD4/6OvSV8MaUlMDYGACePsW1a/D0hI/Pu67Tm1K/bm9ioAn1hfx8mJpWfZTJkJcHC4t3V6H3geBgjByJ5GT4+uKHH9CvH7y8EBSENWuwYMG7rhyBGOgPlvo09DUqCsOH4/FjuLlh61bFu6DERLi6gtze6vHyQvfuWL4cf/2FuXOxdy8mTMCpU5gzBykp77pyb0F9ub2Jgf5Q0cPQ1+holUm16+Ls3h3duuHrrxEcjDFjcPo02rYlBlorOBw8fAgPDzAMuFw8fgx3d+TmwsLi/b509WVkNzHQdYx39Zyui6Gv3t4IDgaPBzOzmknJybUpxOMhMxN8PgCcOYO1axEcjORkfRjo8nJwOLqV0CkUhfx8mJgAgJkZwsLg4oKCApiavt8Gugbv7chushZHnSEqCi1awMICDRtWjeVMToalpT7Uk5MRF1fLZQYFYfJkTJqE5OSaoXZxcsLNm4rjzz6DoyNWrKhlCTnZ2ZgzB76+WLwYz5+jTRsYGaFDB8TG6kROPxQUKALDoLBQcVzP0MXtrRfIWhx1hpkzMWQIHjx46Tlddygd+lrr+Pnpo9uyfj1GjULr1jhxAlZW2LMHffqom7DwxkydipISzJmDf/5B69aYNw/XruG77zB3Li5cqH05/eDiUnXcsqWuVPTm75Kjn9tb9xAXR51Bz8/p9WnoK4DMTAQHo1cvxQUsL8epU3jwAD/8UJsqAgEeP4aTE6Kj0bQpcnJgZoa0NDRt+r72OkUilUk8Xm0K6c3fJae+3N7ExVFn0NtzupwWLdCxIzp2VHgenz/XlVBJieLg6VP8+itu3dKJio0NevdWWOenT7FvH+zta9k6A7CwQFYWAHh44MgRha0pKqplW6ZPeDwloaAA16/XspDe/F1y9HZ76xqGUEc4dYoRCJiuXZmsLIZhmOxspm1bpnVrptbbKCiIcXZmKIrp1o0JCmJMTJiePRkul9mwgQhpYN8+hstlunZlpFJFzK5djIMD8803tSz0bjl+nOHza7/YK1dqv0VeRW83g14gBroukZHBnDrFFBUpPpaVMX/9xSxeXMsqbdsy8+czOTnMrl0Mi8Xs388wDHPyJOPgQIQ0ExPDHDzIyGSKj7/9xpw6VfsqhDdGnzeD7iEGuo7RsiUTGalbCQMDJjqaYRhGJmMMDZnYWIZhmJyc2u+q1z8hOXpoo1cpKWGOH2eGDXu/herT7a0XiA+6jjFiBH7+GRUVOpSoqICNDQBQFLhcsFgAFH+JkDbooY0qKS/H2bMYOxbW1li4EA4O77dQfbq99QIZZlfHuHwZYWH44w/Y2FTdVU+e1LJK5ZADXQ99rX9C0EsbSSS4ehVHjuDkSTg4IDYW//yjk1319CYkp57d3rqHDLOrYygdLlq7I4QoSmVS7d4M9U9Ijh7ayMoKpqYYNQqjRsHTE3Z2CAtTdAxrF70JyalPt7deID3oOob8ZpXJkJUFW1udSBQX66TYD0FIjh7ayMICRUUoK4NMppPy9S8kpz7d3nqB+KDrGBkZ6NMHAgE8PJCYCG9vxMfXsoTehr7WPyE5emijqCicOwcAAwagbVsUFKibifdeCMmpT7e3XiAGuo6xYAEaNkR2Nng8ODujb19Mm6YP3aAgjBpFhLRCP23Upg1+/BEJCdi2DZMmYdgwNGiARYveYyF8ALd3bUN80HUMS0vExsLEBHZ2SE9Hfj4cHevZU9t7zztpI4kEV67g8GHs3/8eC5Hb+zUhPeg6hqPjS5Ohw8Lg5qYrrZIS5OTo481JPRPSZxtVwmajb1+dW2ddC72TS1dernMJ3fGuB2ITXubGDcbSkhk+nDE2ZiZNYqytmX/+qWWJoiJmzhzGwoIBGIDhcBgvL+bXX2tZpV4KydFDG9VX9HDpnj9n/P2Zrl2Zr75isrKY1q0ZimLat2diYmpZSC+wVq1a9a5/IwjVsLLCrFkA0KoVmjXDpk3w8ICBQW1KTJgAAwMcOQJTU7RtiyNH4OSEH35ATg66dydCmtFDG0VHIydHeajd9cH1JiRHD5duzBhkZGDmTEREYNEiTJ2Ko0eRlIQDB+DnV5tCeoH4oOsMUikANGjw0urvRUWKHS5qETMzRETAyQkiEezt8fw5OBzEx8PXt5aXFqt/QnprI70tzqk3Ib1duvq1JCwZB11nMDICAIlEcVDJ8OG1LGRqipgYODkhIwNiMWgaAEJDFQdESA16a6OgIEyZAmNjbN9eyyW/KyG9XTr5krBOTvVkSdh37WMhvEyfPjqX+PlnxtGRmTOHadaMmTWLYRhm6VJGIGBOniRCWqGHNmL0tTinPoUYvVy6+rUkLHFx1D1iY+HsDIbByZMoLsb48bXspANw7Rpu3kTjxhg5EhSFiAg4OsLcvJZV6qWQHF23UWoq7O3VzVp+74Qq0cPtHRuL27cxdqzivPbuhaUlBg6sZRW9QAx0HWPNGnz/PVJSsG8fDh0Ci4W2bfHrr++6WoRq6KGNqm+2rVP0JiTnndze7/PO68RA1zEsLXHlClq1QoMGOHkSJibw8kJ29ruuFqEaemij+mqg9XDpsrOxahUiIuDtjYUL0acPwsPRrh0OH0aDBrUppBfIS8I6hlQKU1NEREAmQ8uWSEmp/WH2ettfuf4JydFDGwG4dUuxxWINan0tUL0JQS+Xrn7tvE560HWM2bNx9SpkMsyejaFD8dlncHHBqVO1KVH/hnDpecdoPbQRRcHZWfkolNpdXUhvQnL0cOnq1zA7YqDrGFIpTp2CRIJhw5CcjOPHMWMGBIJaVtHPyKp6KQS9tFF9dXHo4dK5uuLECXh5gWFw7BhGjACA6Gj06IG0tNoU0gvEQH+QXL2K0FAsWECE6ij11UDrgf37MXs22rXD9euKJ4Off8Z332HyZHz77buu3GtDDHQdw9tbSWRQkN7rQVCNHtqoRw+cO6ePuRV6E5Kjn9ubDLMj6IqQEMUBwyAlBTt3wt8fw4bVpkRsLNzcarodKypw9GgtL1ZQ/4TkkDZ6Y+rTpdMP726ODEELcnOZ1q1ruUyAyc9XHDs4MImJDMMw+fm1vy99/RNSCmmjN+a9vnR6gawHXbdJTkZcnA7LF4n0tJNm/ROqhLTRG1NvLp3OIOOg6xjVnXQSCcLDMWfOu6sNQRmkjd4YculeE2Kg6xhbtrz00dRUJ1MtCG8DaaM3hly614QY6DpGx476UAkNVQw+lUgQEYHsbBQVESFtIW30xtSzS6d7yCiOOkObNiqTQkNrU8jKSmXS8+dESB2kjd6Y+nfp9AIx0HWG8HCVSa1a6bEeBNWQNnpjyKV7I4iBrquUlcHQUFeFl5dj5078+y9iYuDggF69sHAhjI2J0Ouh0zaq39SP21v3EANdl9i7F3/9hYsXAeDAASxdirlz8dVXYLFqU0UkQufOYLMxaxbc3ZGSgr17kZGBu3chFBIhDeinjXr3Vpl06dJ7KYR6d3vrBWKg6wybNmHzZmzcqFjeRSzG5cuYNw8zZuCrr2pT6MsvER2Nc+demm01ZAgcHbFtGxFSh97aaO/equO5c7FyZdUe21OmvJdC9e/21g/vdJoMoRru7szFizUjr1xhGjWqZaEmTZjbt2tGBgUxTZoQIQ3orY2qY2rKJCTosHz9CNW/21svkJmEdYbMTHh61oxs3Lj210hMTYWDQ81IBwekphIhDeitjeof9e/21gvEQNcZWrZEWFjNyAcP0KxZLQs1b65k/bA7d5T8/xChGuitjeof9e/21g/vugtPeMHBg4yLC/PgQVVMZCTj5sb8/HMtCx06xNjZMRERVTFPnzKursy+fURIA3pro5SUqiAUMkFBVR/fU6H6d3vrBWKg6xKbNjGmpoynJ9O3L9OqFcPjMcuX60To228ZY2OmRw9m8mSmTx/G2JhZtIgIaYV+2ghQGd5TIaY+3t66h4ziqGMUFODuXSQlwdYWXl6wsdGVUEICbtxAXBycnNC1qw6XRKh/Qnpro/pH/bu9dQwx0AQCgVBHIS8JCQQCQR+sXbv2xIkTr/UVYqAJBAJBt8hksu7du69evfp1v0gMNIFAIOgWmqavXLny1evPmSTrQesQkUjUoFHjgoJ8PWhRFM1m1+qaBiqQMQylBxlAxoDWixKLpg//tqNfrx66FmIqyigOV9cqABhGph8hCcVmOPpYhIjR180AwIBV1W2VSKQ9B4yUSKWqMstkUpakpKt3Ozc3txkzZqgplsVi0fRrd4iJgdYheXl5memplNCRMjLTtZZn66amZia6VgFQJpXJZPp4sSxjoIcN5aQV5eHn/4yODOvbXg8TGWhGXKx7FYBjzEgr9KAj5ZpJpTI9CFEUpZebDgDYNKgXPwb5hUW3bgWyHDqpzC0ptWM9Hz58uLm5uU4qo4tCCdWhTVxoCw9dq3gNHNnAw1XXKgAKxJIKiT7+J6UyRg9DjMpEheHn/9S1CuE9hWFAsTi0saXKDOXFFeW5Xl5eOqoAMdAEAoGgCgYUBTWuCZo2MNCha5EYaAKBQFAORVEARak10JRMW8/ymjVrXrcCxEATCASCchRONkq1CaboConKV4hvDzHQBAKBoByKoiiKolkq7SRDsww5BrqrADHQBAKBoByGAShKfQ9aostxLMRAEwgEgnLko4gotQZaKtXhWCNioAkEAkE5NEWBomiWSgPNsGhjLkd3FSAGmkAgEJTDgKFAgVI9kI6iS8t0OCeIGGgCgUBQDaVumB1D0W8wgVt76tdiSX/8gS5dIBCgYUNs2aKIDA9HmzavV46BAVTPvn+JM2fQtCn4fPTsiaio11MhEAh1G5qiAFA0S03gcHTYza1HBnrzZixfjmXLEB2NnTuxYQP27tWtYmYmxo3Dxo1IT4ePD0aM0K0cgUDQLzKGASiAVhkoSlwm0V0F6ouBLijAt9/i5En07w97e/Tti40b8ddfL+UJCYG3d81jmQz+/jA3h5UV1q4FgL59IZGgQQOUlCAwEG3awNISY8ciIwMAoqPRowe++w6tWiE4GK1a4ZNPIBDgm2/w6BHya2fhOsmzc0xZwduXk/bs8c+zh3w/uEPAT0ulEuWesuzk+DWf1nzCUBqphqy4J4cXjPxlnM+l7d+oEspLTdg1umPlx4cXj++b0W/nqA4nvpmSl5qgpdDzuCdHvxr92wTfKztXqBLKT0vY7edd+fHRpeP/m93/twm+l7YurSgr1faUAAAPHj5p33+0ZUvfyQtXlFcol+syaHx0nKL+W/f+yfPoWD3cvh+urdAnIyxbdJm84GvVQn7RsQmVH9Oznn8yfpZZM+8ug/yexiW+1nkpRCMetu/Z37Jhy8lzF5aXqxDtNyg6Ju61ig0LDfXp1MHR1nrmtCnl5eXapA7s/4mliUAeRgwZrL1Q547t7W2sZkxVLvRq6h//O9jMo5GNhdmo4cNyc3I0SlCQj4NmqQw0i2NAetAauXsX9vZo27YqZuRIXLyo+YsnT+LKFYSG4tIlfPcd4uLw779gsxEbC7EYQ4ZgzRrExEAgwPjxiq+EhSE1FX/8gV69ULk/QnAwXF1havqW58EUp0uTbjGi529ZDgCZTPrnilkdB46du/d8YU7mjUO/KpFjZCc3LpNUlGuMVFdnmezs93NbfTJ63LaTxTlZd4//plTo0o4V0hdlFmQkX9+z7iP/1ZN++dvCueG1X9dqKXT+h3ktPh41ZkuAKDfrfoCSJySGkV3ZuVIqUQilPrp37/ieHrNWjtx41MDIOOiPrVqeFACpVDZoyjz/CaMiLwekZmSt21FT7trtO1MXrQoOjaiMGTd0wIN//pKHs/u2NXBxbOPZRCuhyXP8J46JvHIqNSNr3Y49rwiFTF20IvhBRPXIIVPn9e7qHXXznLdX61nLvtX+vF6ISgf5TfGfOiHyv8up6RnrtuyoKRp4e+q8RcH3Ql+32BFDBs+Y5X8vLDwtNW3D+h+0SX329OnJs+eC7t4Punt/8/btWgoNGzxo5mz/B2ERaWmpP70i9GpqfHzcvDn+/zt0+F5YeHFx0ZpvtVk+nwEF0LS6oEvqi4FOSoKT05t8USIBw6CwEK1bIyXlpULOnUOPHhgwAKam2LQJ//0HmQwApFJs3YoWLcDnw8oKAM6cwZgx2Poa//yqYEpyQLNB18JvckLEXSOeoG3fIQJzq+5+syKvnXs1T/CpP4SWttpEqiHl0T2OsaBZr0E8M6sOw6c/Dbzwap7w84f5FlU7hBoYcVkcQwMuz8CIy+YYGgm0Wig19dE9Q2N+0x4Djc0s2w2d9uzW36/mibjwklDao3tuHXo4tuggsLRtP3zms1v/aH9eN0PumQj4E0cMtLO2XD532pEzNeVCH0UZc414xlUrL5ubmjR2d5WHgyfO/fbTaq6RoRZCd00EgokjBtlZWy3/fMaR0zUvYOjDmkIRT54WFYu+nDbextLi+yVfbFm1RPvzUojeDjERCiaOGWFnY718wdwjJ87UFI18ZMzl8oxfb7nnW4E3hSbCcRMm2NraLV667OhfRzSmSiSSrKzMzl18Gnl4NPLwcHBw1EYo8OZNExOT8RMm2trZLV66/OiRIxpTg27fbuvl5dWunZOT86QpUx9GRqgouwoG1IvphCrR6eKO9cVA29oqXBCViMU4cABlZcrzV65jOWwYZs3CZ5+hQQMcOPDSAsTJyTh/HnZ2sLODuztYLGRlKbQMXkzuzM3F0KFYswanTmHAgLc/D9q6OcuxE8XS/F+tkdy0JBs3xTKnNq4eeRkpNVbvzEtPvnv2SJ+pCzRGqqcgI9nSpZH82MKlYUFmao11nAsyUyL+Oeoz7ovKGGNTS5/xX/z11Zifx3R+fPVUz1krtBEqzEyxeCFk7tyw8BWhwsyURxePe4+dVxlj6uCa+vBuYWZKRVnpk6unxEX5kjKxlucVm5jSvIlCrnnjhvFJqTUu4Pxp47d9u8TMRPjqd2/fD5dKpR1aN9dKKCFZg9D0Cdu+XVZdKPLJU48GrjOXrG7afYDf3MWmQoGWJ1VNNLF5M0XvvnmTxvFJSTVFZ0/b9sO3Zqavt8h4XGycZ/MW8uNmnp6JCQnVi1WampycbMTljhkxvEXTxpPGj0tLS9VGKD4u1rO54vJ6enomJMRXF1KaOnqM37+XrzIM8+zZ08sXL3bs5K2k3JehKcineqsMNItrqMNx0PXFQHt5ISYGDx9WxVy+jKVLYfiypZO8cOenpCgOkpMxdCji4xEQgD/+wMmTVZltbTFyJNLTkZ6OtDRERsLWFgBYLwZFlpejTx80bYqQEOhsQdg3prSokMPlyY8NjflSiaS8tKQylWGYU5u+6TdzsaExX32kRsqKCzlcRT+Lw+XLpJJycZUQGObKzlW+kxZyjHmVcXmpCXeO7R65/k//v0I8Pxp6afs32giJiwsNjHjVhSpeFrr2y7ddJizgcKsq36hzX/umbf/8fNC+qR+ViQoBSMpV/Ga/Ql5BoYCnkBPy+RUSSXFJifqvvKgIM3/1j9/MU7e/Rk0h/stCIg1Cmdk55y5f92rpeXb/Lnsbq9H+C7XUqhLNL6gSFfArKjSLakN+fp5AoLj+AqGwoqJCVFysPjUrI8PGxmbytGkBp89yOJxxY0ZrVf+8PL5AUL2o4mpCSlMpiqJpOuDE8dbNPS/+++8XX87XqKIw+RSlJpTrcrGk+mKgbWzwxRcYOBDnzyMtDdeu4YsvMGfOS3lMTREejshI5Odj505F5PHj+OwzpKXBzAwsFirbuKgIn3yC8+dx44biDeSoUTVFT52CVIpp05CcjMREJCYqfCB1A65AWF4qkh+XlYpoFqvSjAK4//cxgYW1R4du1b+iNFIjhnxhpemvkAsZVQk9vBzAM7d29epa/Svxd6+7tvWx9WjBMjDsNHp2YmhQeYnmrUaM+MIKseKMKsQimsUyqCb0+MpJnpmVS1ufl75DUd2mL59xKHjKvuutPx1Hs9hGfCUd3kp2Hjhi0aKrRYuu+4+eNjMRFokUckUiEZvN4mv3vB8SFsnn8TzcXdQJ7T9s0byzRfPO+4+eMjMRFhW/LMTTIGRkyOnasd20McMaujr/uHzB3fCHOXmaX1Dv/O2ARYMWFg1a7D901MzUpEq0WMRmszWKaoOpqVlRkaIpi4uK2Gw2j89Xn9rR2/t+eGTffh838vDYvG37vTt3crKzNQqZmZkVFxVVL4pfTUhN6tBhw1PSM2fPmTPgk36az4cBQFEUS03Q6VYv9cVAA1i7FvPnY+VKNGqEadMwZQoWL34pQ+PG8PdHly7o3h2zZysiZ8yAszMaN0abNmjTRvEmcPhwuLjAxAQHDmD2bNjb4+ZN/PnKvhsPHiAsDK6uVSEvT/fnqS3mdk7Pk2Llx9lJcaY2DhRVta1bXFhIVNDVdcO8t0zsK5NK1w3zTokKVxqpUcjExjE3RfGiPzc1XmjtgGpCyREhcXev757ge2D2pzKpdPcE34ynkTKZjKm8rxmGkUm02TxFaOOQmxIvP85LTRBY2VcXSom8E3/3+t5J3f+c+5lMKt07qXvms8icxKfBh7bTLDaLbZD2JNTCxaP6V17Ff8KonMjAnMjAiSMGujs7RMUo5KJiE1wd7Sm1363k5N9X/AZ/oj6P/8TROQ9v5zy8PXHEIHcXp6gXIyWiYuJdHR00CjnZV70koGmapmgWS/Oy8f5TJ+TERubERk4cM8LdxTnqaYxC9Fmsq5OjlmenHjd3t+ioJ/Lj6OhoZxeX6sUqTb17586tmzflkRwOh8ViGRhoXh/O1c09qlpRLi6u1YWUpu7f9/ueX38BYGZuPnXa9MiIiFfHftSAoihQFMViqw4sYyMdrmZXjww0AH9/3LsHkQgxMVi6VOGLaNUKoS/eRG/ejMJChIVh+HAEBQGAQIDTp1FUhJwc7N6tcC4fOoSCAhgbo18/PHoEkQhXrsDdHQAaN8YTRcPjhx/AMC8FCwt9n7JqXFt1KCnMjwq6Kikv++/4760/+kwe/yjwYlFO1qdzV3y+97z/L6embPwfTbP8fzll17CZ0kiNQo7N24uL8uPuXJdWlD04faBJt/7y+Jigy6Lc5z1mLB+37dSYTceHrv2doukxm45buTdxa+cbE3Il9fGD8pLi239ss2vSxpCn2Yvq4NleXJQff/e6tKIs7MzBxr4Kodjgy6Lc577Tlo7ZenLUxqODvv2NoulRG49aujXhWdhE/H0kLuRKYVbq3WO/ePYeqv0F7NapfU5e/tlL18VlZZt2H/QbrJAL+PtyWqa6YTYn/7nau6tm/+YrQtfEZWWb9hzwG6x4mRFw4VJaZpbSr/T27fww6tmxc/8WFYtW/LS9U9uWr+uG7talU05u3tl/LonLyjbt2u03XDG4LeDs32kZma9VVHW6+nbLzck9f+6sWCzetmXzqDF+8vjTJwPS09OUporFpaNGDAu+fbuosHDNqpXenbsITTQ7vn27VRW1dfOmUWPGyONPnQxIT09Tmmppabl+3bqYmGcZ6ek/rv+hTdu2HI4G97GMYRRD7VQAUCW6nOpdvww0oRo0zRqzeueV/dt+GtOdw+X5jla4RAN+WpIW85grMBFa2ggtbQQW1qAgtLRhsQ2URmoUomh6wJKtQYd37p3am8PltRs6VR5/cdvXWfFPjPgmfAtrvoU139yKoii+hTWLbWDh3LDXrBVXdq3aO7V3Xlpivy/XaXNGFE1/snjLnb927Z/e14Br3HbIFHn8lR0rsuOjjPgmPHNrnrk1z8yaoiieuTWLbWDEN/lozprA3388+c3khp37vpaBZrHogD1bVm7c5dKpr4BvvMRfITdpwYqwRypnjUY8eSouK3O0s1GVQbnQb9tWbtzp0rG3gGe8ZM7UF0JfqxIyMjQM+G3ruh17nDr0ioh6enDrD0qzqRVlBfxvz8p1G11adq9HEDMAACAASURBVBLw+Uu+8FeIzlkQFvnodUurXuyR4yfWrl7VuIGbQMBf+JXiKXb61CnhYWFKU7v6dluybPm4MaM9Grg9jY7ee+CAlkJHTwR8u2plI3dXgUCwaLFiHMv0KZPDQ8OUpg749LPxEyf27tmjVfNmkZERB/7QvBclRYECRdG0msCidbjlFaWHfTk/WFJSUpycnFhOXfSwaazfNLJp7JtQJir8bYLv5pWLPp/sp2st+cQz3asAHGOw9bHMTjnXTMrmas731lC14n/RDiM2q1IsJ6/QwbOzWZdZqjJLS/JMsq4/u3tZR5UhiyURCASCcuQbXqlZD5qiqbJysmA/gUAg6B1K7uVQ/QKWolkcsqs3gUAg6B95D1rdfG6K1qkfjhhoAoFAUIF8U0I1Lg6KrtDl7AdioAkEAkE5FK0YB60yA4vF1eWu3mSYHYFAIChHvhw0RdFqQrmKQU0Mw3z++ee9e/ceMGBATrWlTSUSydixY729vX19fRMSEtRXgBhoAoFAUInaeSoUKEqVC/rq1atZWVmXLl0aOnToxo0bK+PPnTtnYGAQFBQ0Y8aM9evXq1cnBppAIBCUo3mqN80yZCEvL+/VWeOBgYEdO3YE0KlTp8DAwMp4oVBYWFgok8ny8/OFQnUrw4D4oAkEAkEV8t6xulkyMklqTOSMGTM8PT1XrlxZPSU3N7dly5YAXFxccnNzK+O7d+/+9ddfe3h4PH/+/PHjx+orQHrQBAKBoByGYeR7XqkKtIGRbcPWR48erWGdAZibmycnJwNISkoyNzevjN+wYUPv3r1jYmL+/vvvCRMmqK8AMdAEAoGgHMWu3mphsZT3r318fO7duwfg/v37Xbp0qYzPzs62sbEBYGVllaNpX0Ti4tA5TVt4OLZ+vRWW34DnGdkJMW+yc+jrYu7eSMrS4RYSlRhy2BLdr69dUVIKABQNXS55I4fiChm2ka5VAJTJqApK56cDgIIBo9PlkF9QIC4X6XJGdSUUhYYWvKrFOBjFVBXVX4CqURy9evU6c+bM4MGDAezZs+fx48fjxo27f//+okWLxo0bd/jwYbFYvF3TBozEQOuchk3cvXp00rXKn3v+in2aoGsVAA0lXGNzaz0IGXENKFrnK+RISsUAQNHQvUVjuKYU9zW2qnljxKUVYok+7CaHYbH0sthajqg8p0SHq3pWx8XMmFV541EUKNCq70Oapjhs5X4IiqK2bdtW+dHS0vL+/fsArKys/vlH270xiYEmvB4SXW7wUx0Zw7DUdV0IqqGoyt2adAoDRm33shaF3g2Vy0GryqDrC00MNIFAICiHogAKqmd6g6JRISVTvQkEAkHvUAAFdS4Oiqa4HB1aUWKgCQQCQTlSxUtC1W4cHe/qTQw0gUAgKIdW7HmlMgNFER80gUAgvAsUw6BVrwdNUbShAXFxEAgEgt5hGFCUuul8NFBaTlwcBAKB8I6gVMwVlCex2TocaEgMNIFAICiHohQTvVVnoFi6nE5FDDSBQCAoR/YWU71rhfdnsaQjR+DjAxMTeHpiwQKUlCjiraxAUS+F2Niqb3l7Izq66uOvv8LdHebm8PODSPRS+SJRVQkcDtq3R0iIysqEh6NNGwA4fx6DB9faORIIhLoETVHycdCqAkVThmwdLhLwnhjotWsxfz7mzEFkJPbswaNH8PZGxYu5+deuIT+/Kri5AcDVq5gyBcHBVYVcv461a/HbbwgLA5+PJUuUCCUmIj8fCQno0wcjR+r+xAgEQt3lxZZX6vZUkTIfeA86MRHr1uHffzFqFJyd0bkzLlwAi4UdOxQZBAKYmFQF+ZiY0FAYG4PHqyrnxg0MGoSePeHsjJUrceSIEi15Cfb2mD8fSUkoKUFICLy9FanVj2vw5AmcnBAUpM0JpT17/PPsId8P7hDw01KpRPkSMNnJ8Ws+baNN5NsgeXaOKSt4+3JK0mOe7J4Ttn5YwumNjFRSI7WiOPfZn1+H/TAkau8X4pxUeWT2/QuRWyeEfj/w6YGvxDkpWgoVpz4L2zojeOXAZ0d/fFWovDDn0d4lQSsGhO+YU/pcUWbW/X/v/eAX9M2AJwdXSEoKX+u8HkQ+bv/xMEvPTpPnLy+vUN5SXT4bHR0bX/mxn980XoM28jBwkv9rydVUf/CgXbv2FhaWkydPfnXDjrchMjysr693Uxe7L2dPr1BR8qcfdYt99vQNCg8PC+3euaObvY3/jKmvVvvV1J93bLM1F1YPd4K1+j969ihi1uCPBrf3+GnJ55KKmkK3r/wzqV+XAa1cF44fkhz3TB6Z+zxz6dRRn7VtMHfExykJsa8UqQQKNR/RXwqALmd6vxcG+vp1dOiAFi2qYmgas2bh33/VfWvBAmzfDjOzqpjGjXHtGuLiIBLh99+RnV3lJ6mBSISDB9G1K4yNtaphair698euXSrNdzVkMumfK2Z1HDh27t7zhTmZNw79+moehpGd3Lisxj2nNPKNYYrTpUm3GNHzWiiKkcUcXmnd4TNP/90VhTnpgTV/+WKPrBa6t/Wcu5fv1Czp3DYAZXnpSX/vcv3sy+bz9nOtXZPO71BW8CtCMtnj/cvtugxuu+D3soLnyVf/rJHhyYFvTBt5tfvqf0JXz5iATQDEuekxAVua+K1ou+B3aVlp4sX92p+XVCodNMnff6Jf5LWzqRmZ67bvrpHh2n8hUxd+HfwgvHrk09j4C3/8+uBiwIOLATu/X6G93KvqAwcO8vf3f/gwMjU1dd26dW9c1KslTxg1dPL0WddDQtPT0rZt/LFGhls3r8+fM/P+3TtvVvjoYYOnzZwd/CAsPS1t00/rNaaO8hsbGHxPHv46ccrNzb1lq9YahWRS6Tczxw0cO2XvhcDszPRDv2ytnpqX/fyHRf4zl64+evthc6+O334+VR6/YtYEry7dDlwM9mzTYcuKRRpVFGtxsCg1gcv5wF0ccXFwd68Z2aABYmIUx927w8xMERwdVZYzciS6dkWTJrCzQ14eAJSW1szj4qIoZ9EiaNrPUUFeHvr1g68vPv1Um+wJEXeNeIK2fYcIzK26+82KvHbu1TzBp/4QWtpqE/nGMCU5oNmga+EtcXFCBMuIZ9G6jwHf3M53dN7Da9VTSzPjpWWlNt5DDHhmDr0mO308CwBtYESzOTSHyzIwotgcNlegjVBBXDjbiGfTrh9HaOHUa9zzsCvVU0XpcdKyUgff4QZ8M9ePp7kPnAOgMOEh39GD79TY0NTatkN/UbpWnSY5N4PvmQj5E0cOtrO2Wj5v5pFT52tkCH342JjL5RlzK2MkEmnm8xyfDl6NG7g1buDmaGejvVwNbty4aWJiMmnSRDs7u+XLlx8+rOyB740I+i9QKDQZOXa8ja3tF4uWnDx+tEaGh+HhXK6xsTFP6dfV81/gTaGJid/4CTa2dgsXLz1x9IjGVDMz80YeHvJw5NAfO37dY8Tlqii+ivC7t3kCYb+ho82tbPxmz796NqB66pOw+w2aeHbs9pExjz/Wf35CTHRxYUFc9OMSUfGwSbPMLK2mLlzu//V3ms/nxUxvNaHsQ5/q7eT0kitZTmoqHBwUx4cOoWVLxbH6WZk//4xt28AweP4c27ej2j40CgIDIRRCKsXNm+jXD3de7kQwymZ13riBBQuweze+/x729hrPJjctycbNQ35s4+qRl5Ei31anMkNeevLds0fGfb/78a2L6iPfBtq6OQCmIOntiyrLS+dau8mPjaxdy/IyUW0VytLMeCNLx8Rz24oTwrnWbo79ZgAw4Js5fDQl6rd5oCg2V+A5Z682QuKcNGNbxU81z9ZNnJuhmEgAABClx3KtnGJObCqIC+PZurt9OhuAdZuPrNt8BIYpzU7Ne3pX6OKp/XnFJiY1b6JoqeaNG8Un1Wyp+TMmATj9z+XKmKS0NK6R4dBpnz+OjmnfpuWPXy90sH1DGx0bG9uiRXOFevPm8fHxNdTfmMT4uCbNFNehSTPP5MSEGiXPnDsPwD/nzrxB4fHxcZ6eimo39fRMTHipcPWpd4KDpFKpV7v22gilJyW4eTSVH7t5NMlISapeVJvOXT3bKsp5Enbf1sGJLzQJvnbJya3B5m8Whof859a46exlazSqMC82VFGVgZL3sXXG+9CD9vHBrVtV/WU5e/eiZ0/Fsb09XFwUwdlZZTkREVi2DAYG4HBw6xZatlRizZ2d4eICd3dMnAgvL1y7BgCSF77OFGWu0r59sWEDhg3D119rczalRYUcrqJvYmjMl0ok5aVVnhaGYU5t+qbfzMWGxnz1kXUHSWkxy1DR5WEZGjMyibRcXJlaIcoriA7m2TVsOGaNgcAi7tj3AMQ5Kek3DzWZuqXNstOWbT9OPL1RedE1hYrYhgqnE8vImJFKpOVVz0AVxXm5T27zHT08J6/jCC2i/vwWUHR+siNu3N8wIS/6jkO3EdqfV15+oeDFOwwhn18hkRSLVPjEXpCRlW1rbTlj7MizB3/hGBiMmjlfe7ma6nl5AoHiwUIoFFZUVBQXF79xadUpyMvnvyiZLxBUVFSIRLVTMoD8vDw+X1G4QCCsqKgQVau2mlSGYZYuWvDVUq3+iQAUFRQY8xT/DsZ8gURSUVpSNS6La8wzMbcAcPvKP9/PnynvLOflZAVdu+jRvOV3ew5ZWNuu+WKaRhW5A5pWi6GBDq3o+2CgmzSBvz9690ZAAJKTcecOhg5FejoWvXAhFRWhoKAqqHiZA0dH7NiBgADEx2P1akyfrog/cQJpaYpjeQm5uTh9GsHBaNYMpqYID0dkJPLzsXOnkmLlj2Nr1+L4cYSGajwbrkBYXqq4k8pKRTSLxeFWebrv/31MYGHt0eGlLbKURtYd2Fy+tExhKGXlpRTNYnGqNnai2Ry+SwtLr08Mze0d+0wtSXsqKS0siA42adSe59CEZnPsu48rjH0gLdNg+wCwuQLJi2zSshKKZrE4Vc/CNJsjdGtp23GAkYWDa/+ZxcnRla8ELVt177TylL3PkId7vlIvsXPfnxbNOlo067j/r5NmpsKiF2Mxi0QiNpvF52l4J9G5XZuH18593NO3cQO3nd+vCAmNyM7N03heSjEzMysqKlKoFxWx2Ww+v3Z+nk3MTItflCwqLmaz2Txerf3wm5qZFRcrCi8uLmKz2bxq1VaTev/uHb5A0LBRIy2FBCYmJS9+V0qKi1ksNvdln0xRQd6qOZP+2Llp9a4DnXr0AcAxNGrZ3rv/yPEOLm4zFq+KjggrzH/D1qmOuEKHLo73wUAD+PFHrFiB9evRrBnGjYOdHe7erXqD16MHTE2rwtGaPjUF5uY4cADz5sHXFyNGVBnoiROrDKuLC0xNYWGB2bOxejW6dkXjxvD3R5cu6N4ds2errKG9PRYswHzNPSZzO6fnSQpPaHZSnKmNQ/UHqLiwkKigq+uGeW+Z2Fcmla4b5p0SFa40UqOQ3jA0sxNnK1wl4uxkjqlN9Yc+A6FVVVZKPmSJZhhZNX8Rw8ik0GKskpGFXWmWYt/F0qxkQzPb6s9AhqZVG3FRCgchnXnnQnrQGQBsY4Ftx09F6bGvjv2ojv8kv5zHITmPQyaOHOzu7BT1LE4eHxUT5+rooNHDEBIacSP4rvyYY2DAommDN11Jx93d/cmTKIV6VJSrq2ut+DcAuLi6PXsxOSDmabSjs0ttlQzA1dUtOkpR7WfR0c4uLxWuJvXs6VPDR47WXsjOyTUpVjHIJDkuxtbRqbqQpKJ88aQRLg08dh7/x6N5K3mklW2VB5KiKYqmadWrIMlhAIpSOQiapimKogxYH3gPWs6kSQgJQVERoqOxYweEQkX88+dgmJeCn1/Vt5KT0bhx1cfBg5GcjORkrFpVFVlUhP79weO9VEhqKhYuVGTYvBmFhQgLw/DhioF0rVopbHr//jh5UpFt5UqFS0Qtrq06lBTmRwVdlZSX/Xf899YffSaPfxR4sSgn69O5Kz7fe97/l1NTNv6Ppln+v5yya9hMaeQbXEIdwXdtKSkpzI8OlknKM2+fsGipcD3lPblVUZQjbNC2NCs+79FNaXlp2tWDPMemLCO+SaMOeU/+K056KC0rSb2yj+/UjGWkuRNn4t66oqQw9/FtmaQ89eZR67YfyeOzI2+WF+aYerQryYjPDr8uLStJ/Hef0KUZm8tn80ySr/5Rmp1SXpiTcvVPvoMHxdLWYnbzbp+Tl3/24jVxWdmmX/f7DVG8BA64cDEtM0vpV8Ri8dApc/+7+6CwqHjFT9u6dGhrItDq/eerdO/eLScn58yZs2KxeNOmTX7V7+q3w9vHNy8v5+KFc2Vi8S/btwx9YRbPnzmVmZ7+loX7+HbLzc35+/y5MrF4x9bNI0aNkcefOXUyIz1dVSqAs6dP9vyot/ZCrTp0LszPu33ln/KysmO/7+r12TB5fOC/53KyMv679LdUKv1kxLis9LTM1JTM1BRGJmvn0yM++vGNC6dLRMX7t/zQrLUXX2iiXkVu8imaUhdq7+ftVd4fA11foGnWmNU7r+zf9tOY7hwuz3f0DHl8wE9L0mIecwUmQksboaWNwMIaFISWNiy2gdLId3sW1aEouuGolWnXDkZuHksbcm19RsnjE05tLMmIpdmcBqNWpgceidw4pjQz3m3IVwC41q4un85LPLMlcpOfOCfVdchirYRoutmENYn/7rv73UiWobFTT4XNevbX+uLUZzSb03TCmuSrf95ZO0KUEesxahkAC88uNu36Rf78xf2fJojS45r4aeviBMBisQJ+375ywzaX9j0FfN6SuYpHrklfLgt7+ETpV7p5d/j6i1mjZs536dAzKjbuf9trjmB7LfWTJwNWrlzp7OwiEAiWLlU2r+pNS9536NiP363xataQLxB8Pl/hKvxi5rTIiLC3L/zPoye+/3aVZyN3vkDw5SJFy86ePiUiPFRV6qOHkWKx2L7ytb8W0CzWtz8fOLB1/WjfVsY8/piZ8+TxPy6eG/M48tnjiNgnD/16tK0MRYUFHEPDb38+eOiXLaN8WsZFPV66YZdGFcVTHqUuSKQ6XBGaYpSOTCDUBikpKU5OToPmr/H6eLiutfS2q7ert6/Q2k4PQhwjNkuXD49yJKVFwSsHbl699PMp43StBTMH/ezqnS+WiCt0OX3iBQYsmq37ndcBxOeKcvW1q3cXNwujFxt15xYUNezU23OyykHoZflZ4qBDEVdP66gy78MwOwKBQHgXMIqBHOqG2UllOuzjEhcHgUAgKIeCfC0OdUGmwgnBMMznn3/eu3fvAQMG5OTkVE/asGFDjx49OnXqlKJ05G41iIEmEAgEFcgXhFb3hhBGBsqnel+9ejUrK+vSpUtDhw7duLFqpH9ERMTp06evXr26bNmyHTs0LHJAXBwEAoGgCka9iwMUVaZiPejAwMCOHTsC6NSp0++//14Zf+HChR49elAU9emnn/r4+KiXJz1oAoFAUAFDgVI3zI6RSZ7HRy1ZsmT37ppraeXm5rq4uABwcXHJzc2tjM/IyIiNje3bt2+/fv0SExPV65MeNIFAIChHvqs3rXqkCtvAwMLWYfjw4WbVF84EAJibmycnJwNISkoyr7bsD4/HKy0tPXjw4IMHD6ZOnXr//n01FSA9aAKBQFAOo2kpO4qmWcYCLy8v91dW3PTx8bl37x6A+/fvd+nSpTK+Y8eOpqamLBbLzMxM4yQXYqAJBAJBJRQFFk2pDBTYKkbr9+rVy8zMbPDgwcePH1+4cOHjx4+9vLwADBgwoKCgoGvXrmPGjNm2bZt6deLiIBAIBOXI+7cs1d1cmlLZyaUoqrr9tbS0lHszaJr+5ZdftKwAMdAEAoGgEo0TVSp0OVGFGGgCgUBQCUVRqpwYANgs2pCtQ0cxMdA6p6xCJirT4YqxclhCC9tmQs353pry4qKM7FrYyVAjZs4NaZahrlWkYjEAMDLIdN5GFEXpQQWAAQWWLjfKq6Jq5xzdwqIooZE+jBXDVK6QpIAC1Cw3QlHQ6VRvYqB1TrmUKdXlkt5yrBo2NirVxz9/QtDNwqy3XZRSGxgDE0O+ha5VZGUvDDSj86vHSCsomT6WITQyYIPSx/v/MolMpkvzVAnfkG2gy63/qlO9u8yAAQVatYuDpqDT8ycGup7AYrMBPd3B+kGmr39I/aHFpgS1AlWzF6hDIf2glyXzlCAfSsdWva4/myIuDgKBQHg3MBQFSvXvA0VT5SqmetcKxEATCASCKihoHGany+49MdAEAoGgHOrFRBVVGVg0ZaDGfr81xEATCASCcuSjVNT0kWkKFbrc8ooYaAKBQFAORYECxVI3ioM2IC4OAoFA0D+MwsWhMgOL1u1YFmKgCQQCQQUMoGmqt0yXgyeJgSYQCATlyNcUVbNzOYumOGQcNIFAILwDGFBqF+ynaUh02YWuF+tBHzkCHx+YmMDTEwsWoKREEW9lVXN57djYqm95eyM6uurjr7/C3R3m5vDzg0j0UvkiUVUJHA7at0dIyEsZahRFIBDqBQwgn+qtMuh4NuX7b6DXrsX8+ZgzB5GR2LMHjx7B2xsVFYrUa9eQn18V3NwA4OpVTJmC4OCqQq5fx9q1+O03hIWBz8eSJUqEEhORn4+EBPTpg5EjFZGvFkUgEOoLNAUKYFOUqsCiKI6ad4hvXwHdFa0PEhOxbh3+/RejRsHZGZ0748IFsFio3MxcIICJSVWQz6kPDYWxMXi8qnJu3MCgQejZE87OWLkSR44o0ZKXYG+P+fORlKTop79alBZkxT05vGDkL+N8Lm3/RiqpUJonLzVh1+iOlR8fXjy+b0a/naM6nPhmSl5qgpZCJekxT3bPCVs/LOH0RkYqqZFaUZz77M+vw34YErX3C3FOqjwy+/6FyK0TQr8f+PTAV+KclNc6LzVInp1jygrevpzy7LjUgK8SD0x4fn0nI3vpjAojzyXsHV09iDOjARQ9uZR8eHbigQnPr25lJGWvJffg4ZP2/UdbtvSdvHBFeYXyluoyaHx0XELlx/Ss7E8m+Jt5dukyaPzTOA1bglYJhYa16+htYWM3eer08vJybVLPnDvXrEVrgZllrz79orR4gAt98KBD+/bWlpZTJk9+VUJp6tmzZ5t7epoKhb179YqOitLyXMLDQrt6d3S2s5k1feqrQkpTB33a39pMKA8jhw3WUuhxZPiw3r6dGrssnze7oqKmkJzR/T+Kj30mPz64e1cbF5vqIfRuiNJvVSIfxUHTlJqg03HQ77mBvn4dHTqgRYuqGJrGrFn4919131qwANu3o/omj40b49o1xMVBJMLvvyM7u8pPUgORCAcPomtXGBsrL0oTjEx29vu5rT4ZPW7byeKcrLvHf1OSh5Fd2rFC+uKeK8hIvr5n3Uf+qyf98reFc8Nrv67VSoiRxRxead3hM0//3RWFOemBNX91Yo+sFrq39Zy7l+/ULOncNgBleelJf+9y/ezL5vP2c61dk87vUFbw68EUp0uTbjGi2lihlJFl/PuD0PNjx+FbpCW5+aEB1RP5Ht0dhm6UB5t+ywyEtoYWbuK0R/kPjlt2m+UwdCNlYJQb8of2alKpbNCUef4TRkVeDkjNyFq3Y2+NDNdu35m6aFVwaET1yCHTvujt0ynqxhnvdq1mLdOqpaRS6cDBQ/1nz3oYFpqalrZu/Y8aUzMzs8ZPnLLhxx/SkuJ9unQeOXqsRonBgwb5+/uHR0ampab+sG6dxtTMzMyJ48f/tGFDcmpqFx+fUaNGaXkuI4cOnjFr9p3QsLS0tI0/rtcm9dnT6BOnz/4Xcu+/kHubtmp110mlUv/xo/ymTD97MyQzI2331o01MoTcuvn1l3PC79+tjBk4YnTA1Vvy8PMfR51c3Zq2aKlBRuGDhpqgZp7h2/OeG+i4OLyyVyMaNEBMjOK4e3eYmSmCo6PKckaORNeuaNIEdnbIywOA0tKaeVxcFOUsWoT1618tQ0tSHt3jGAua9RrEM7PqMHz608ALr+YJP3+Yb2FT+dHAiMviGBpweQZGXDbH0Ehgoo1QcUIEy4hn0bqPAd/cznd03sNr1VNLM+OlZaU23kMMeGYOvSY7fTwLAG1gRLM5NIfLMjCi2Bw2V/DGp1kJU5IDmg26Fl5Hl6Y/ojnGgsY9WMZmpm2GimJuVU+lDfkGpvbyUPzsumU3f4rNEac/MnbtwLVvweZbmnkNF8XeUlX4q9wMuWci4E8cMdDO2nL53GlHzvxdI0PooyhjrhHPmFsZE/HkaVGx6Mtp42wsLb5f/PmWVV9pI3TjZqCJicmkCePt7GyXL11y+MhRjanBd0JatWzxycf9BALB18uWPnr8OD9f3QPKzZs3hSYmEyZOtLOzW7Z8+ZGXnxGVpoaEhLRs1erjjz8WCATLv/768aNH+fn5Gs/lVuBNoYnJ2PETbG3tvlqy9NjRIxpTJRLJ86yszl18Gnl4NPLwcHBw0Oai3Qv6jy8QDh411srGduaXi86fPF4jw+PIcC6Xy5V3pAAAJqZmbg0aycPpY4e/27LTyIgLtVAUQIFF0eqCLo3oez6Kw8lJif83NRWVbXzoEFq++JFUs4EuReHnn7FtGxgGz59j+3ZU2yZdQWAghEJIpbh5E/364c4deHi8QZULMpItXRrJjy1cGhZkpoJhqtetIDMl4p+jg77ZFRt8WR5jbGrpM/6Lv74aQ1G0kUA4fudZbYTK8tK51m7yYyNr17K8zOrrq5dmxhtZOiae21acEM61dnPsNwOAAd/M4aMpUb/NA0WxuQLPOTX7jG8Abd0cAFOQ9PZFSQozOeYu8mOOubOkKFPpivHizGjIZIbWDQEYmDqIHhyXFGayjE2Loq5KxUWMpIxia7UPQGxiSvMmipZq3rhhfFIqwzDVh8TOnzYewOmLVb98kVHPPBq4zly65kbwveaNG21asVArobi4Fs09FUKenvEJCdWFlKb26tGjcydveWRwyB1XFxdTU3U/23GxsS2aN5cfezZvnhAfX11CaWrPnj29GX2IqQAAIABJREFUvRUSIcHBLq6upqamGs8lPi7O01NRVDNPz8SXz0VpakpyspER12/U8CePH3u1a7923Xp7e3uNQkkJ8R5Nm8mPGzVplpKUWKN1Js2aC+Dy3+de/W7o3RCpVNqijZdGFfntpWZtbYqCLheze9970D4+uHWrqr8sZ+9e9OypOLa3h4uLIjg7qywnIgLLlsHAABwObt1Cy5ZKrLmzM1xc4O6OiRPh5YVr15QVpJmy4kIOV/GrzuHyZVJJubiaO4Vhruxc5TtpIce4yq+dl5pw59jukev/9P8rxPOjoZe2f6ONkKS0mGWo6CCwDI0ZmURaLq5MrRDlFUQH8+waNhyzxkBgEXfsewDinJT0m4eaTN3SZtlpy7YfJ56u+dj4bpGVFdMGRvJjyoDLyKSyCvEruZjc2/tMvYbLP/AadDa0a5py9POk/02VlYsAMBLlzspXySsoFLx4uyDk8yskkmJVjq8XZGbnnLt8w6tFs7P7dtjbWI3216oHnZeXLxAoHlaEQkFFRUVxcbH6VD6fb2VlCeDMuXN+4yZs2bRBk0Qev6oQ4SsSSlL5fL6VlRWAs2fPjhs7dvOWLdqcS35+VVECgbCiokJUTUhpamZmho2tzeQp046fPMPhcCb4jdZGqLAgj8dXFMXnCyQVFSU1Bl+pgGGYdd8smT1/sTaZ5YslsWlKVWDRlJpR0m/Pe96DbtIE/v7o3RsbN6J9e6SnY/16pKdj0SJFhqIiFFR79DM2hoGyLS0cHbFjB9q1Q5s2WL0aX3yhiD9xAt7eMDEBoChHKkVgIIKDsWrVm1XZkC8sL1X8n1eUimgWi2NU9RT28HIAz9za1auruKjqcTL+7nXXtj62Hi0AdBo9++cxnctLijnGfPVCbC6/pEzhqJGVl1I0i8Uxqkyl2Ry+SwtLr08AOPaZGvr9IElpYUF0sEmj9jyHJgDsu48L+2GItKyEZWistHz9QxvyKy0yUyGmaFalva6kLOsZbcA1MKnsglGWPtMtOk8BGFlpYeHDC7SRuuu288CRFRt2Atj4zUIzE2HRi//5IpGIzWbxjTVcCiNDw64dvKaNGQrgx+VfCpt65+QVWJgp6dvu2PXzilWrAWz86UczM9OioiKFUFExm83m86sqqSo1Nzdv+sxZScnJJ08c82rbRn3FzMzMiqsKKXpFQnlqbm7uzOnTk5KTTwQEtPXS3N8EYGpaVVRxcRGbzeZVE1Ka2rGT990XfvzNW7fbW1vk5GRbWFiqFxKamImKFUWJRMUsNttYu3f1EQ/u8fh81wYNtcksn0ioZi0ONUm1wnvegwbw449YsQLr16NZM4wbBzs73L2Lyn+kHj1galoVjh5VXoi5OQ4cwLx58PXFiBGYPl0RP3EiQkMVxy4uMDWFhQVmz8bq1eja9c3qa2LjmJsSJz/OTY0XWjtU760nR4TE3b2+e4LvgdmfyqTS3RN8M55GymQypnJjIYZhZBKG0fzi2NDMTpytcCyIs5M5pjbVvQEGQquqrBQFiqIommFkqCqZYWRSvW0Cog1sgU1FnmJgSUV+Kptv9ap/QxQfwmtU1TTluYl5dw5RNIui2eKMJxxzF/VLJ/hPGJUTGZgTGThxxEB3Z4eomHh5fFRsgqujvZopv3Kc7KreHNA0TVM0S4WHcs7sWblZGblZGZMmjHd3c3sSpRiGERUd7eriUl1IaWp5eXnfT/o3bdok+L9AjdYZgJu7+5MXwzCio6JcXF2rSyhNLS8v/7hfvyZNm94OCtLSOgNwdXOLjlYU9TQ62vnlc1Gaeu/unVuBN+WRBhwOi8UyYGveGMzJxTXumeKyxD176uDkrLF15Fw6f/bToSO0PB3qRSdaTZBp8c/4xrz/BhrApEkICUFREaKjsWMHhC/2Tn3+HAzzUvDzq/pWcjIaN676OHgwkpORnPxS17ioCP37g8d7qZDUVCx82bdYoyi1ODZvLy7Kj7tzXVpR9uD0gSbd+svjY4Iui3Kf95ixfNy2U2M2HR+69neKpsdsOm7l3sStnW9MyJXUxw/KS4pv/7HNrkkbQ57m13d815aSksL86GCZpDzz9gmLlgq3T96TWxVFOcIGbUuz4vMe3ZSWl6ZdPchzbMoy4ps06pD35L/ipIfSspLUK/v4Ts1YavubeoZr7yktKypJvMtIKwoizvAb+crjRfHB0pJc+XFJfAjXsVXlV9g8i8JHf4viQyRFWXn3jwma9tZerlun9jl5+WcvXReXlW3afdBvsKKlAv6+nJapfFBKb1/vh9HPjp27WCQSrdiws1PblqZCzS3VvZtvTk7OmXPnxGLxps1b/MYonvFPnDyVlp6uNPXUmbNSqXTqlMnJKSmJSUmJSUkytfPZunXrlpuTc/bsWbFYvHnTJr8X/wgBAQFpaWlKU0+fPi2VSqdMnZqcnJyYmJiYmKheQk5X3265OTkXzp0Ti8Xbt2weOXqMPP70qZPp6elKU0tLS8eMGB4cdLuosHDt6lXenbsITTS/Bm/f2Sc/L+/avxfKysT7f9n+6VDF1ISL585kZajbM/PShTNduvfSWP4LKIoCm0WrCiwWGQddv6BoesCSrUGHd+6d2pvD5bUbOlUef3Hb11nxT4z4JnwLa76FNd/ciqIovoU1i21g4dyw16wVV3at2ju1d15aYr8v16mXUAhRdMNRK9OuHYzcPJY25Nr6KIZJJZzaWJIRS7M5DUatTA88ErlxTGlmvNuQrwBwrV1dPp2XeGZL5CY/cU6q6xCt/HT6g6Jt+izOu/dX8p/TKQOuSZsh8ujs6zvKsuMBlOcmyqTlbF7VVrO0Id+qx5zc27+nn/mG16DzaxloFosO2LNl5cZdLp36CvjGS/ynyOMnLVgR9kj5oGAjQ8OAPVvW7fjNqX2fiCfPDm75Tjsh1skTx1auWuPs3lAgECxdrHDQTZoyLTQ0TGlqaGhoWHiEW8PGlSEvT90QCxaLdSIgYNXKlW4uLnyBYPGLqVhTJk0KCwtTmhr64EF4WFhDd/fKkCcf4KTpXA4fO/HdmlVNG7rzBYIFixS30MxpUyLCQpWmdvXttnjZsgl+o5s0cn8aHbVn334tL9r2/Ye2/fhdzzbNeDzB9Hnz5fHL5s188jBC1beiHz8qLyuzsdP8ErIS+Y5XaoKqXb0Zhvn888979+49YMCAnJycGqmFhYUuLi6a1bV5WCa8GSkpKU5OTr1mr2ree4iutR7G5uQXvd4sjDdDb7t6W3i055pa61pFViZKPDBh88pFn0/205z7LTFzoPTyRCKj2YxedvUul8hUmafaJaekvLRCT5sIN7TkG7zoFBcUiTr1GTDlhz2qMudnpgX+sf3GmWOvJl25cmXPnj1HjhzZt2/fs2fPvv/+++qp8+bN27t3b/VXtUohPWgCgUBQjmI1O3UuDtpAxZ7fgYGBHTt2BNCpU6fAwMDqSffu3SsqKnJyctJYAWKgCQQCQTkMw6h/SSitKAsNutm7d+/ly5fX+G5ubq7cieHi4pKbm1sZL5VKly1btl67yW7v+TA7AoFA0BnykSG06vEhHEOjVt5dL51SMjzM3Nw8OTkZQFJSknm1iW87duwYMWKEfIy5RkgPmkAgEFRCgWLRKgNNU6rMt4+Pz7179wDcv3+/S5culfEPHjw4duzYxx9/nJyc3L9/f/XqpAdNIBAIqmAoCmoWfaYAmYrXpL169Tpz5szgwYOB/7N33mFRHG8cf/cK5Qr16OUAUUDAjg3FgsSeoEQsqGAvWLF3TTQaY8UWozHRaDSiYEFFUSxYQGlSBKT3fpS7o13Z3x93woFcAT1M/M3nmYdndmb2/e7sHu/Ozs7OwJkzZ969ezdr1qzo6Ojz58+LCtjZ2d25c0e2PHLQCAQCIQ1RF4fUbCJg0mazwzDM39+/aZPBYERHR0sWSE5OliuPHDQCgUBIBZM3zZpShxkiB41AIBBt82HJK6kFCBgo9VMS5KARCASibURzcciYEYmAYdLGQX8WkINGIBAIKeCAgaxZmDAMBKgFjUAgEF8KGeOgRavKKg/koJUODrhSJyQUIcRxsgpR2SoAQKbpaKootObWJyKor63OV3Sh0g6D8+oAADCCrGUzPhMY8Wv7d+u0mXxEX/R1Bhi0qBMmXjRWWnECQeoojs/C1/aL+ReC47i0kZKfERVVEq9T/lk0Ta0a6ztjkuiK92/qqkqVLiNoWmBF6R4AFwo7y8t0kg5grfyZsiAQCBjWSTcDyfayKCbj1o2B1NnsPgvIQX8lyHgK+8xCRCLAv2gW//8UnTVzZMtVLpVHZ90HOk+oTWnZK6oodVEV5KARCARCKhgmsw8aMKX2jSEHjUAgEG0jeuSRPYpDgXVmOg5y0AgEAtE2mGiyJJnjoNFLQgQCgfgCoC8JEQgE4l/Kh1W9pc9mp+SXschBIxAIhFQwABmdGEQ0DhqBQCC+DBjIHqaBASj1IwfkoBEIBKJtPozikNXFodRODuSgEQgEom3kzmZHVPI3Yl/FmoRXrsCQIaCpCfb2sGYN1NaK0/X0Wq/Bm5HRvNegQZCa2rx5+jRYWYGODnh5AZfbwj6X22xBRQWcnCAyUpx16xbY2QGNBiNHQorSJ45AIBCdCi5rSW9RUOoojv++g969G/z8YNkySEiAM2cgKQkGDQIeT5z7+DFUVTUHS0sAgLAwmDcPIiKajTx5Art3w9mzEBcHNBps3NiGUE4OVFVBdjZ88w1MnQoAUFICs2bBwYNQVARDhoCnp7LrikAgOhUMAICIyQpKnc/uP+6gc3Jg7164fx+mTQNzcxg8GO7eBSIRjh8XF6DTQVOzOYim1o6NBQoFqNRmO0+fgrs7jBwJ5uawYwdcudKGlsiCsTH4+UFuLtTWQkQE9OwJ48YBnQ7btkFSElRVKXLUZZnJV9dPP+vt8ujEdgGf12aZqsLs37wGNW0mhV77a+n4s94uoUc38RrqFDs7wClIizu6KGLHd2lX9+MCfqvcxpqKpN83vto+4e3xZXVl+aLE0uj7Ufu8Xm2bkHxhO7+2RkGhxvLMgsD1Oee9y56cwIUthGoSgrN/ny4Z6ktSAYCdHJp3eWnOee+ysKM4v0FBIbnw04LxhurPZS0mMdlpnCfD0Xnumq2NvLavlLO7V2pGdtPmmJmLqF37icJ3c5crKhQb12/AIF0Do7nzFzY2NiqSeys4uLtjL7o2w/WbMSmSz4JSiI2J6e/kpM9gzJs792MJGblDnJ3fK2C/ibdxsUMHDTA3MliycP7HQm3mFhcXTf5uookBw3XY0PS0NAWF3sW/new2tL+N+aaVS3i81kIipo5zzUpvNlhWUrxg+uR+1iZTx7lmZ6TLlRB1cRAwTFrAMAwjoBa0NJ48gf79wdGxOYVAgCVL4P59WXutWQPHjoG2dnOKjQ08fgyZmcDlwrlzUF7e3E/SCi4XLlyAoUOBQgFXV7h+XZweEQEWFqClJfeQcaHwzr6VjmOnzTgSyGWVRgf+3kYZXPjoxA4BX/ybK0iKirp2ZsSSHVMPXiWrUV5dPCpXRST07s8tRs6T+qw511Bdlhd2qVWB5PPbtLr27bf+Lw0L+/TAQwBQzypKDzxi67W9z5pzgoa6nAd/KiIEuLD4/j4N+7GmU44IallVsYGSmbRuw008DoqCwZjNZA1DVV3L+sKkqphrjGFLTDwOYmQ1VuRFhYRkHwWnSJD7HOeWfbopEQKB0H3uMl+fGQmPbhQUl+49fqZVgccvI+ev2x4REy+Z+D4j++6FUzEh12JCrp3Ys1UxIcF3kzx8ly5JjIstKCzc+/N+ubklJaWzfeYd2L+vMDdriPPgqdNnypWY5O7u6+v7NiGhsKBg3969iuQ+fvx4wfz5kZLPmgrUZarHpEVLlr6OjSssLDy4/2dFcqdP+X6kq2tsQtKAgYNWLfdVUGjJ7Kkz5y668+x1SVHh6SMHWxWIeP50y2rft9FvJBN9vacPHjby/qvYPk4DdqxbJV8GxwDkdXEo81Pv/7iDzswEK6vWiV26QPqHe+Pw4aCtLQ6mplLtTJ0KQ4eCrS0YGUFlJQBA3UetVCZTbGfdOvj5ZwAAGg309AAAbt2CGTPgqEJ+syApSpVCsxvxHUWb0c9jQdrzex+Xib97maZr0LRZmBRl2X+EqWN/OsPQacritOchighVZ74lqVEN+o1R0dA1c51VFvdIMpdblCloqDNxmUKmaVuMXWD13TIAqMlOpJl2o5nZqGrpG/Yfzy3KkGK7BXVFSQQVCt1mBJGirdXbg5v+XDKXoEojaxmLAiftCWOYL0ZSqS9Kolj0Vzd2JNEY2n2ncDOeSzOuOHhtBRBIQPhs772fRb7RpNN9PN2N9PW2rFh05ebdVgViE1Mo6mpUinpTCp8vKCmvGNK/r00XC5suFqZGBqAAT5+Fa2pqzvGebWRkuGXTxstXrsrNjXgd2bOH47ixY+h0+tbNm5LevauqkvXc8OzZMw1NTW8fHyMjo81btlxp+YwoLTcuNpZCoVAlnzXl8Tz8mYam5szZ3oaGRus3bgq4ekVubmJCAofDXrZilb6+wc4fd/984JAiQm9ePafTNSZPn6lnYLhk9frgoIBWBZIT4tXU1dUplKaU1HeJXA5nzuJlunr6flt3btnzM8gFwzFMPNhZRlDkgDvGf3wUh5kZfHx7LygAExNx/O+/oUcPcVz2lCenToG/P+A4lJXBsWOgo9O6THg4aGiAQADPnsGYMfD6NXTrBiwWLFgAublw4wb07avIIdeU5Osyu4riOubWNSUFrSaHrCnJT3pwbcKW45mRYpeqZWKRfu1MTUm+upZuctiNenYVv6GepKomW6i+opBiKL57UQ0t61nFkkLcogx1PbP064eqM+OohlaWE5cCgH7vUfq9RwGO15UXVL5/o8G0V6RG/JoSFR2mKK6iY85nlwDgH88QWV+SCkKhqr41AJC1TLgx1/g1JUSKFjslTFDPxvkNGElVETlpEPQdAACvzv0UI5JkZOc52IqvlIONdVZuAY7jkoOq/BZ6A8DN+2FNKbmFRepqqh4LV717n+7Uy3H/ljUmhvryhTIzHR3Ep9rB3j4rO1tSqM1c1xEjBg8U94BFRL62YDK1tGStopCZkeHo4CCK2zs4ZGdlSUpIy13t5wcAt27elFuFJrIyM+3txaa629vntKxLm7lJiQldu3ZbuWxp+LOn9vYOe39p3RZuk9zsrG524tPSza57fm5Oq6szZ8lyAHh0705TSuq7JEvrrtvXrnz9Mrybnf2mH/eCQsgaSYcpeZrq/3gLesgQeP68ub0s4vffYeRIcdzYGJhMcTA3l2onPh42bwYyGVRU4Plz6NGjDW9ubg5MJlhZgY8P9O0Ljx9DYyN88w3Y2UFkpILeGQDqOTVkNXGTREWdJhTwefUS3Sk4/vjXH5y916io05rSug4ebWzX59IK9z/mj2rg1gAAv1F+py2/jk1SFTcfiGoUXMAXNDY/FvA4lazklzTTbvZz96po6KZc+gFA/CxXHv80+oB3Zeprk2EKvfYUNnAIZPHdAiOr40KBkFf/USmc9fIPrb5TRBvULoNVjezyr67I/Wu+sJELADi/7T7EL0hldQ2dJr5SGjQaj8/ncKV0fH2guKzcUI+xyGvK7T9PqpDJ05asUUiosopOp4uFNOg8Ho/D4cjOpdFoenoMALgVHOw1y/vIoQPyJCppzUY0PpKQldsuqqqaTdHpGjwejythqs3c0tKSe3fv9Ord51rQLUNjY5+ZMxQRqqmqotLE/yM0Gp3P49W2Gnz1ERVlpY8f3HPo2eu3S9f0DQ1XL/CRqyJqaBCkBxmNZxzHV6xY4ebmNmHChIqKiqZ0Ho/n6enZv39/JyenqKgo2QfwH3fQtrbg6wtubhAYCHl58Po1eHhAURGsWycuwGZDdXVzkPKeB0xN4fhxCAyErCzYtQsWLhSnX78OhYXiuMgCiwU3b0JEBHTvDjdugEAACxZAXh7k5EBOjiIzD6rRNHj14l8Sr55LIBLJas1PYe8eBVG19Zh9hrTYB8OGLdyy6O+IeX886TVxFoFIUqNpyBUiqdP5DWKHImioxQhEokrzwziBpKJh2cNwwAQ1XROL8Ys5ealNrwQZPYcP3HHDeMjkxDPr5aoAAEGV1uSRcV49RiA2+esmGkrTCGR1sqZxU5UYQxYy5/5t7v2HpuNEjEAkqNHgX8CJPy/rOgzWdRj859Ub2poabI74SrG5XBKJSKNSZO8+uG+vxLCbY0cOtelicWLP1si4+HJWZZslj588paNvqKNv+Mf5C9raWmw2WyzE5pBIJBqt+WxIy2WxKr/3nLZ7z96g6wETxo+TfWDa2tqcZiPsjyRk5bYLLa1mUxwOm0QiUSVMtZmrqqrmPGTonHnzrbp02f3TvpjoKBarog3TLdHQ0mpy/VwOh0giUeR1xaioqvYb6Ow5a465pdX6HbsT4mKqKlmyd8FAvOSV1IBh0sZBh4WFlZaWhoaGenh4HDzY/Fhw7do1Go32+vXr06dPr1ixQvYB/McdNADs3w/bt8PPP0P37jBrFhgZwZs30NTxNGIEaGk1h6tX2zaiowPnz8PKleDiAp6ezQ7axwdiY8VxJhO0tEBXF5YuhV27YOhQiImBuDiwsGgOlW3/N0qiYWDCys8SxSsLsul6xpKt9fyE11lvnvw+Z/il5d8KBYLf5wwvSUuoyHkf8fcxApFEJJELk2N1md0UWS9DTdeorjRHFK8rzVPVNpTcS1Wr+dEbE7/vIJS8vlv06hYAkCh0wwETuUUZH4/9+BgS3YBXKR4EwqsqINH0Pm5YcLMiqV2HNm02snIqX/+NEYgYgVRfnKyiw/yii2Y04+szvSLxZUXiSx9PdyumWUp6pig9JT3LwtRE7kdjkbHxTyPEbSIVMplIIJLJ5DZLLlu6hFVazCotnuM928rSMjlFPEwiJTXVgsmUFGozt7GxcfS48XZ2thEvwvv26S23XpZWVskfxumnpqQwLSwkJWTntgsLS8vUVLGp96mp5i3r0mauqalZUwECgUAgEIhE+atrmjEtMtLEpyUz/b2JmbncYzYyaX4LJXawBDlCuLiw1FEcBAyTNgw6PDx8wIABADBw4MDw8PCmdBsbm/Xr1wOArq6u3GP+7ztoAJgzByIjgc2G1FQ4fhw0PrQuy8oAx1sEL6/mvfLywMameXPSJMjLg7w82LmzOZHNhvHjgUptYaSgANauBQDYt6+1fV1duQdrYu9Uz67KevNEwGuIu3XBxmW8KD0j4iGXVeayYNOMo0HTDl51/+EsRiBMO3iVYWlL1TWIv3clM/JRTWnBm4Bf7d08FDkrmla9eLU1rHcvhfzGgmdX9fuMEqWXJzxrrKnQ6tavtjir/O0TQUNtzv0/NJjdSeo0ElUzL+xiXXl+Y01Fftglmkk3RdY5VTe2FzSwa3Pe4AJedfwtWlcXUTo3K0JQK26h1GZFqpv2bNqFRNWtSbrHzYrks0srowPodm6K1KiTGTbQqaKy6nbo4/qGhkNnzntNmiBKD7wbWljS9mKJ9fUNHgtWvoiKreFwth845uzUR5Muvyk6fJhLRUXFreDg+vr6Q4ePeM2YLkq/HnSjsKiozdwbt24LBIL58+bm5efn5Obm5OYKZT69DRs2jFVRcfv27fr6+sOHDnl9+EcIDAwsLCyUltsBhroMY1VU3A0Orq+vP3bk8NTp4v6KmzeCioqK2swdOWpUUmJi4PVrHDZ7966dTv0HaGrKHw3Vf/DQqkpW2P27DQ31504d+/b7qaL0B8E3S4uL2tzFefjI98lJ924Gcjmcoz/v7tXPSUNTzvLHH74klBoEvIbkxLcbN248c6b1IB8Wi8VkMgGAyWSyWM1N9T59+tja2kZFRXl4eGzbtk32AXwVDvo/BUYgjNtw5PU/J/9cOJqsTukzeZ4o/dHx7eVZKWo0TaqOPlVHn6qtj2EYVUefSCKr0TRHLfsx/Nz+oG1zrQePVtBBYwRCd+8fc+7/8WbPVKIqxWyk+L8u7Z+fOQVpBJKKnfePeWGXXu/25BZndJu2GQB07Z0N+o1JOLUq+hdvblGmrZdCo8QAIxh8s6Ey6p+8Swsxsrpm78mi5PInxxvKswCgkZUjFDSSqM13L4IqTW/EMtbLc0W3tlG7DP53OmgikRB41n/HwRPMAW50KmXjsvmi9DlrtsYltf3V6LBBTltXLpq2ZC2zv1tKRtZf/gq9hiISiUHXA3bs/NHcyppOp2/aIO6gmzNvQWxsXJu5sbGxcW/jLa1tmkJlpawx+EQi8Xpg4M4dOyyZTBqdvuHDp1jz5syJi4uTltsBiETi5YDre37caWdtRaPT16zbIEpfvGBefFxsm7lqamqXA64d3L+vWxeLxMSEM3+cV1DoxJ+X/X/eM7yXHZVGW7RS3N2/ccXidwnxbe6iqqp24vzl00cPuvTslvoucf/J1i5VCpgMVFRUTc3Mp0yZMmLEiFa76ejo5OXlAUBubq5Oy0EHW7duXbNmzblz58aMGSNHu9MWTv8/JD8/38zMbMSS7d1dJytbK7OIza6T0sP+WakoYtdzO0Wos1b15iVcOrxzw4q5coYSfwa0jTDVdgxZ6zBCAglX6kp5H2jkC5W6pnUTlXW8Op6gE4QAoIsujUQU9zxwOLXjvpt04i8pXaMARYX5vx74KfBK648MAODhw4fnz5//66+/Ll26lJCQsG/fPlH6vXv3Lly4cPHiRUV6clALGoFAIKSA4aJFY6UGkHondHV11dbWnjRp0rVr19auXfvu3bu+ffsCwKNHjyIiIvr169e7d+9vv/1Wtv5/fBw0AoFAKBnZX1BIe72NYZi/v3/TJoPBiI6OBoADBw4cOCBnWGQTyEEjEAiENMSfekvNVvLgI+SgEQgEQhriLg5p2QQMU+psdshBIxAIRNuInK88B4xW9UYgEIgvAQZSvxUE8SgL1IJGIBCILwRB+mA3AkFW7qdMRvqiAAAgAElEQVSDHDQCgUC0DQ4gr5ND6qfenwXkoBEIBKJtPqyoIrUAQamvCJGDRiAQCNnIGcWhzKF2yEEjEAhE2+CYeE1CaQUwJc+WgRy0EhHdWmMuH0u5dU7ZWjwBrsyVdySE+EJipwhhPD5d+TMR4DheCbB5//FdR9tYHPLzoqquJv3D4M8JgaDcjtEmVEiEHfsODB81WtlC3EZBo6BT5uLAMaHEekCi5VJk9UBjgFrQ/1UMDQ3XLVtUxSrvBC0BTlDq2mjNQgIBsVNmcOELcZJSX5ADAEBDY+OFa7fqcbUGgvypYj8RMzOmKrUz1iVQI3fGFRIKBIkPg2LfxvcY3Hoit88OTyDkd8qsTADQ+t6GYTLuqsq+4SIHrUSIROLP29bj9ezOECOQOqd1BkLeRz9h5YDJXlHo81BZXXPh2i2CjjVBr7uytazHjjXt+tEax0qArkZSJcmfKe0T4Tc2JD4MUrbKFwYTvSeU1cWhVH3koBEIBEIKuHjCfmkQMeVOx4EcNAKBQEgBE/2R1QutzC+9kYNGIBAIGchsI2NKntAOOWgEAoGQCgayxmlgaDY7BAKB+CKIXC9B+vgoAgFT6htz5KARCASibUTDoGX2QKNPvREIBOILgWHy5uJAfdAIBALR+WDy2siYUqfrRw4agUAgZCOjDxojKLcF3SnfnnUaFy+CszPQ6WBtDUeOiBPfvoXevdtnh0wGBT/8P38erKxAQwMmT4aKivapIBCI/wKYzKBUviIHffgwbNkCmzdDaiqcOAEHDsDvSp7+JjMTliyBf/6BxERgs2HHDuXKIRCIzgcDTDbKFP9aHHR1NfzwAwQFwfjxYGwMo0fDwYPwzz8tykRGwqBBreNCIfj6go4O6OnB7t0AAKNHA58PXbpAbS2Eh0Pv3sBgwMyZUFwMAJCaCiNGwJ490LMnvHgB/fqBkxOYm8OCBRAf365DjklMdho/ndHDZe7a7Y08XptlnN1np2Zmi+JHf79E7TZAMryMfquQUMI7p7HfM+wHzvXbIlXo2+mpGVlNm2O8FlC79BaF7+b4KqlGAFBUWj7O21fb3tnZffb7zJx2CI3zZDg6z12zVbqQV2pGs9CYmYuoXfuJwndzlysoJAN+WjDeUP3pdiqyU4K3zLy8cOSL07uE/NZ1Cd237KLPYFEIO+gnSsyLeXZj3feX5g69v2dxdWG2gkIlGckX/TxPeDmH+G8VfCQkglWQ7T+1f9Nm/P1rZxaM9vd0urplLqtAUaGUxLezxo1wdbTatWYZj9fYKvdZ6L0pIwa42JoumfZddkaaZNZc929yMtIVVAGA90nx8791Hd+n6971yz8WevEwZOY3g79xZK70mpT7QWiNj+eo7qaisGnhTEVUMAAihkkNoNxx0F+Lg37zBoyNoU+f5pSpU+HBA/k7BgXBo0cQGwuhobBnD2Rmwv37QCJBRgbU18PkyfDjj5CeDnQ6zJ4t3iUuDgoK4OJFmDkTnjwBHIf37+H+/WbvrwACgdB93kpf72kJDwMLikv3Hm/d2H/88vX8dTsjYpud/iyPCTEh/4jC7T/8uzBNe9vbKiAkcJ/j6+vjlfD4dkFxyd5jv7UWehE5f+3WiJgWvv59Rtbdi6djHgTGPAg88dN2JdUIACYvWOU2ZGDK01uD+vVcsnm3okJzl/n6zEh4dKOguHTv8TMfCUXOX7c9IqaF0PuM7LsXTsWEXIsJuXZiz1ZFhKSBc4oEuc9xbtmnGBGbEgrDDvjZfuPpvv9qbWVp/M0/WhWoKcoZteHYxL2XJ+69PHDORgCoq2aFn9zWz2u158n7Bja9nvpvVFDoxp5lvcbP8Dl+g1NRGhnQ+qQBAI4L7/tvE3zwdFXFeWG//TR62Q/zfwthMK0f/arQ1REKBH5zZ3j6zP/n0auy4sI/jh+WzGWVl21fuXjVtt0hUSm9nAZuWjJHlP7m5bMf1y1PiIlSRKJJaNOCmZNnz78Q8ryspPjiySOSuZXlZbvXLF22+YcbEUk9+g3YvnyeKD0vM/2XP/45F/zkXPATvx/2KyIkXvFKVh+H/NeEOI6vWLHCzc1twoQJFe3pC/1aHHRuLpiZdWRHPh9wHGpqoFcvyM9vYSQ4GEaMgAkTQEsLDh2CFy9AKAQAEAjg6FFwdAQMAwIBAgLA1hZCQmDtWsVln0VGadJpPp7fGekztixfcOXWvVYFYpNSKOpqVIp6U4qOlqaNlYUoXLgefPaXXepqqvKFIqI0NWg+UycZ6ettWbn4yo07rYUS31HU1SWF+HxBSVnFkP59bbpY2nSxNDUyUFKN4pPfsznc1QtmGTB0f9qw4sjO9YoJvdGk03083Y309basWHTl5t2PatRaiM8XlJSLamRh08VCwRpJA6+tAAIJCJ/hBXtxcrQKhWY97Ft1LUYP93lZr0JaCAkEddUVBja9NY2YmkZMio4+AJSnJ+iYdzPt5UxWo/SYNL8qP7OxVv50iXmJb1QpdAdXd6q23kDPRSnPWp80AIgN/pvOaD4zZFV1ooqqCoVCVlMnqaip0TQVqVF05AsaXWOipxdD32DeirX3b16TzE2IedOtu4PzSDcKjTZv5drM9ynsmmoASE1MUFOnqFMoikiIiHv9kkrXGPf9dF19A29fv4e3AyVzk+KiutjZDxw+ikKleS9fk52WyqmpFgj4rPLSnv0GmltZm1tZ6xkaK6hFwKQGTLFvvcPCwkpLS0NDQz08PA4ePKh4Nb+WURyGhuIuiCbq6+Gff2DatLbLN02Y+f33UFQE334LBAL4+sKyZc1l8vLgzh0wMhJvEolQWirWIpObi3l6wqhRcOYMfPMNxMYqeLwZOfkOtl1FcQcb66zcAhzHJbuz/BbMBoCbDx5/vO/L6LcCgaB/LwfFhHIdbLt9EOqalZvfWmjRHAC4GfKwKSW3sFBdTdVjwYp3qelOvXvs37rWxFC+R+tAjRJS0rp1sVi86cenEVEONl0PbVfoDpeRnSdHaKE3ANy8HyZRoyJ1NVWPhavevU936uW4f8saE0N9RbTahKDvAAB4dW6HLTTBLs3XMrMWxbVMrTmlhYDjTf/wnIpioora48NrqwoyGV3s+81YSdHWN7J30uvaQ1SgLD2BpmekQqHLFaoqzmMwxSeNwbSuLimQFAKA6uL8tyFXJ28/lfZK/EugajNcvFdfWjsDwwhqdM25p24rUqOC7GxrW/HErV1s7ApzcyWvjpPzsB59xV0oCTFRRqbmdA1NAJi50BcAntxv3XqQJZSTbWVjJ4pbdrMtymsh1Hewi0MfsdC72GhDEzOahmZhXo6qmvqWpT7Z71PsevZZsmmnnoFR29YlkdnPrGAfdHh4+IABAwBg4MCB5861Y/mOr6UF3bcvpKdDYmJzysOHsGkTqLZsY/L54kh+vjiSlwceHpCVBYGBcPEiBEnMb2toCFOnQlERFBVBYSEkJIChIQAA8cNMu7//DqdOAQDo6MCiRfD2LTS27giTRmV1DZ1KFcU1aDQen8+prVVkRxzH/Xbt37ZykaJCVR8JceUIFZeWG+ozFs2cevvCrypk8rTFfgoJtb9GJeUVwQ+f9nXsfvuP48YGetN9FWpBV1bX0GntrFFZuaEeY5HXlNt/nlQhk6ctWaOIUCfQyGWT1cXNRhV1qlDA5zXUNeXWVVWoa+l2c/VwXXeUQCSLejNIahQ1DW0AyIt59uz4lv6zFbqrNXBqVNSpH4RoQgG/sV7ipOH4gxM7h89dp0KhNqWxCrIjrp6e8culFVdfO7pNDjm6TRGhmuoqKk28KAGVRufzeXVcblMuhUrV1mUAwLPQe1uXL1iza68iNtuEXVNFaRKi0/l8Xl1ts5A6haqlowsALx6G7Fq1aMX2nwCAVVaqo6f/3XTvn3+/TCKr7Fg2TxEhDHCC9NBYX/fo0SM3N7cNGzbIMMJisZhMJgAwmUwWi6V4Nb+WFrSBAaxaBd99B/7+0Ls3pKbCqlUtmsMAoKUFb99CQgKYmcGJE+LEa9fg0iW4cwe0tYFIBA5HnM5mw7hxsHkzPH0KvXrBkSMQEgKvXrUwyGCAry+MGgU0Ghw+DH37goqKjGM8cf7K9gMnAODgtrXamhrsDz9cNpdLIhFpij3fRcYl0KjUblZMWUJ/XNr+iz8AHNyxUVvrIyGqHKHB/XonPg4Wm/ppu5atUzmrkqGj/dlrpKaqOrR/3wUzPABg/5bVGnaDKiqrdbXbeJQ+8efl7QeOAcDB7eu1NTXYnHbWqG+vxLCbYlN7tmp1HyCtRp2MCpXOqxM7Sl49FyMSyarNPTP63Xq47w8QxQfO3fj3PJcGdpUqXauBU/Pq7G5ORfFIv4O6lnaKCKnSNBrrxCetsY5LIBJV1JpPWnzodaqOnmXfoXXsqqbEzNdPLPsMNerWAwAGz/A9Pn1QYy1HhSJnRRgNTS3uh38iLpdDJJHUqVTJAjVVlbs3rCouyD9w9qKdYy9FDr5N6BpatR+EajkcIpGkTmkttH+zX2lh/k+nL9g49AQAhz5Of91/Icr1+2H/mJ6W1ZUsTW2dDh8DAKipqbm6up4/f152MR0dnby8PADIzc3V0WmH4tfSggaA3bvBzw927ICuXWHBApg3D1rd02xswNcXnJ1h+HBYulScuGgRmJuDjQ307g29e4vfBE6ZAkwmaGrC+fOwdCkYG8OzZ3DpUmvF776DOXPAxQVsbCA+Hi5fln2Avt7TKhLCKxLCfTy/szI3SUkXj5pIyci2MDVWcLhO0L1HXpPGyRGa41XxLrLiXaTP1ElW5mYpaZliofRMC1MTuUKRsfFPI96I4ipkMpFAIJPbvpF/Yo3MJPqCCQQCASMQpaym5eszvSLxZUXiSx9PdyumWUp6U42yFK5RlESNiGTJTqovB13fpLpQfNKqC7NpDGPJbofy9MSS5BhRnEgkYwQiRiQJ+bzQfb6aJhbjf/hTQe8MAFoGZqx88Ulj5Wdp6JtICuXFv854/eTkLJdziycIBYKTs1yK3scLhUIcF4pL4LhQwFdkdVQTJjM7/b0onpOeZmxqLnl1eLzGZV4eltbd/rwV+ineGQCMzZk5GWKh3Mw0Q1OzVkJrvKdYWHf9NfC+yDsDwLu46LjIl6I4WYVMJBBJJAVaqDgALpQZ5J+WIUOGREVFAUB0dLSzs7Pi1fyKHDQA+PpCVBRwuZCeDps2ifsievZs7ho+fBhqaiAuDqZMETeH6XS4eRPYbKiogN9+E3cu//03VFcDhQJjxkBSEnC58OgRWFkBANjYQHJys+KPP0JREdTUQEgIWFsrfqTDBjpVVFbdDn1S39Bw6LcLXpPGi9ID7z0sLJE1PCAoJMxtaDuGiwwb5FRRWXX7weP6hoZDp//0mjxRLHT3QWFJaZu71NfXe8xb/uJNTA2bs/0Xf+f+fTTp8ns5O1AjN5dBialpAcEP2Fzu9gMnBvbpoaWhuNDj+oaGQ2fOe02a8KFGodJr1OCxYOWLqNgaDmf7gWPOTn006Z2xNqBcDO36NbCr82KeCXiNSXcvdhkyVpSe8zqstrKMz2sIO7y29P1bXh039topA5teKhRabtQTXCjoOmISt6KEU17EKS9qdqPSMXN0qmNXZbx+zG9siLp5vvtw8UlLexnKYZW6Lt4y5/jN2UeuTf3pD4xAmH3kmr6VnVU/l/SIR/nvYhprOeF/HTW2661KlX91+g4cUlXJehZ6r7Gh4eKZ42MnTRGlh929XVZS/PT+XYFQ4D5jdklRQVFBXlFBnlAo/+DbpPcA55rKyhcPQxobGq6cPfmNu1joaUhweUnx8wf3hELBhKmzSosKigvyigvyhEJhQ0P9liXeCdGvuRz22UP7HPsNoNI1FBLDcekBFBnF4erqqq2tPWnSpGvXrq1tz2iCr8tB/3cgEgmBZ47sOHiSOXA0nUbZ6CvuDpuzZntcUoq0veKT39c3NLRrEAKRSAw8d2zHAX+m00g6jbpx+UKx0OrNcYnJbe4ybFD/rauWTFvsx+w/MiUj869jCo1G6kCN1FRVA88c2Xv8rJnTN/HJaReO7FFU6Kz/joMnmAPc6FTKxmXzPwhtlSY0bJDT1pWLpi1Zy+zvlpKR9Zd/x7s+Py8YgTDS70BcwK8By8eR1aiO34qHnT0/vZOVnWpo17fnpPlP/TcGLB9XXZg9dOmPAFCRncLKeX995cSm0MiVP4oDIxDcN/u/+PvEb/PcVNQp/b8Xn7R7R7eWZqao0TRpuvo0XX2qjh6GYTRdfSKJzGBauy3dEXp8x+m5oyoLc8b57VOkRgQi8eDZi78e3Dt+gAOFSvdZtlqUvmvN0tSk+JSEt++TEr4d1LMpsKurZBuUIbTn9IXfj+zzcO5BodJmLlkpSv9p3bK0dwnvk+LT3iV6uvRpCpyaqt4DnL2XrdmxfN73zj1zM9K2HTqlmBQOQoHUgCv0yTGGYf7+/kFBQUFBQQwGQ/FqYoo8tiA6DF6aiRaN7SCdtWgsw9GZaDKgExaNdXH/2haNPTql37JNO7w/OEfl0Zmrevc311Elif+Vamtrp3hMDr4pdW3c3Ly87bt+PH/hLyUdzNfykhCBQCCUhIwWiaijQ2kgB41AIBDSwAEHWf0YCrwA+BSQg0YgEAgZyGwjoxY0AoFAfDlwWc1k1IJGIBCILwcOMsYCCoVKXVMFOWgEAoGQDi67mazQhyodBjloBAKBkILI98rq4lDu4D/koBEIBEIGOMgYgq3k0dnIQSMQCIQUMLkfc6NRHAgEAvFFwDHxp97SEAqU+rkrctAIBAIhG/SS8CsFFwo6Z4oMjKKBk9Tll/t0IV4t/jlWe5IvJBR0hhCfBAAmXSwMerVjjsCOUcvmxj17Jb/cJ2Pfr4caTenT9QkaGwEAAyAofcYUqBcIG3jKHXEsAgcQtHC4uAip5XGljrJDDlrJYAQirvwZfwAAV9fC1Kjyy306fKpSlzFuBhdiyr+3YTwMAAyZpnYD+sgt/Ik8u3Ev/8Pc3EpFy8CQwZS/XuUnImjkAQCGAUGxqcw/hQa+kMtTaN64TwdrvYW6OL5iOmeGuc4EI3TibHaIjsDj8b70IXxmOqVFIB1ZP3hckfm4Owxy0AgEAiEd2R+q4EKl3jyQg0YgEAgZoHHQCAQC8W9ENA5a1peESl3zBDloBAKBkIHMyZJQFwcCgUB8SWTPB61MkINGIBAIKYi+85Y9HzQaxYFAIBBfAEy0ZoqMIdhCUOYwcOSgEQgEQiYyuziU+pIQfQuAQCAQ0sDFXRzSAybrQ3B8xYoVbm5uEyZMqKioaErn8Xienp79+/d3cnKKioqSIf+1O2gMg4KC5s3ffwc3N3FcTw8wrEXIyIARI1onSloYNAhSU1tL2NqCnh7w+cqvDAKB6GQwwAGEQllBehdHWFhYaWlpaGioh4fHwYMHm9KvXbtGo9Fev359+vTpFStWyJD/2h20bB4/hqqq5mBpCcHB4ri5eXPc2BjCwmDePIiIaG0hNhZqakBLC8LCvkQFEAiEMhFNhoTLCrj0QXjh4eEDBgwAgIEDB4aHhzel29jYrF+/HgB0dXUxmV3Y/98Omk4HTc3mQCAAldocp9HEcQyD2FigUID60WxEly/DzJkwbRpcudJe8ZjEZKdxngxH57lrtjZKmTzB2d0rNSO7abOotGzc7CXa3Qc5u3u9z8xpr2IL9ZiYfv2cdHUZc+fObWxs/BRTzTZj4/oNGKRrYDR3/sKPbbaZeys4uLtjL7o2w/WbMSkfP518OaGK7JTgLTMvLxz54vQuIb/11Qndt+yiz2BRCDvoJ0rMi3l2Y933l+YOvb9ncXVhtoJ1kQs/LRhvqP50O1U5qY9+8L693C3q3I8f1+j5oZU3FruIwkv/taLE+ury54dX3Vrm+njPfE5JbgdEkxPezhg7fJi95Q4/Xx6v9ZV6+uDepOH9B3czWej5bXZ6WnuNZyYnrJ86xntI9xPbVvM/+g96citg6dhBswbZ/LJ6PruqUpT45smDld8N8xrQdef8KQVZ6fI1RN5Tev9GQ0P926TkjRs3njt37uO9WSwWk8kEACaTyWKxmtL79Olja2sbFRXl4eGxbds2Gfr/3w5acdasgWPHQFu7RSKOwz//wKxZMG0aBAVBe9ycQCB0n7vM12dGwqMbBcWle4+faVXg8cvI+eu2R8TESyZOnr/SbeiglGfBg/r2WrL5h45WBgQCwXffufv6+iYmJhQUFOzdu7fDplrYnOThu3RJYlxsQWHh3p/3y80tKSmd7TPvwP59hblZQ5wHT50+818ihAuFYQf8bL/xdN9/tbayNP7mH60K1BTljNpwbOLeyxP3Xh44ZyMA1FWzwk9u6+e12vPkfQObXk/9NypSFzmHwSkS5D7HuWWfwZRQ+PLY2i4jp7jtvlxXWZZy53yrAuziHOfVR1x3/uW686/es9aLEl8dW29gP2D0TwG61o4xF35ur6hQIFg1Z/pUnwXXH0eUFhf9fuyQZG5FWemWlYvWbN8TGpPaq//A9Yt92mdcKNi33Gfs9DlHgp6wSosDz/pL5pbk5/z240a/X04dDgyr43L+OXkAAKoqyvw3r/Beu+NsWKxt7/4H1y5SSAnHcaFAWiARiIYG+qNGjRK1lEX8+eefU6ZM2bNnj46OTl5eHgDk5ubq6OhIWt26deuaNWvOnTs3ZswYGeL/Bw66e3fQ1haH5ctbZA0f3pxlatpuyy9egK4uODqCvT2YmMD9+4rv+izyjSad7uPpbqSvt2XFois377YqEJuYQlFXo1Kap3iOT37P5nBXL5htwND9aeOqIzs77gKePn2mqak5Z46PkZHRli1bLl9ud/O/DZvPwjU1Ned4zzYyMtyyaePlK1fl5ka8juzZw3Hc2DF0On3r5k1J795VVclvKnaCUHFytAqFZj3sW3UtRg/3eVmvQiRzcYGgrrrCwKa3phFT04hJ0dEHgPL0BB3zbqa9nMlqlB6T5lflZzbWshU+eW2D11YAgQSfY1Ls8tQYsjrNYsgENU2G3cQ5eZEPWggJBQ01LEbXXnRDJt2Qqa6tDwDV+em8+tqu30xX1dBx8Fjac7pfe0WjIl7QNDS+m+rF0DdYsHJtyI3rkrkJMVE23R2GjHSj0mgLV63LeJ/CrmnHg0JSVASFrjHCfaq2nr7HwpXP796QzE2JfdPFvqe1Qy+GkckoD6+c98kAkBYfY9Gte5+hI9WptCmLVudlvOeya+TINC0aKyUQCLiBrvaoUaPs7e2bdvLx8QkICNiyZcuQIUNE7wCjo6OdnZ2bCty7dy8jIyMsLKxHjx6y9f8Phtndvw9GRuL4lSvw8GFz1t9/Q9MJ6sBgxsuXISVFbLyqCv75ByZOVHDXjOw8B9uuoriDjXVWbgGO45K9UX4LvQHg5v3mru2E5Pfdulgs3rjracQbB9tuh7avb/cBN6lnZDg6OojVHRyysrJaqXfEZmamo4P4N+pgb5+VnS1ps81c1xEjBg8UT5MfEfnagsnU0tL8NwixS/O1zKxFcS1Ta05pIeB40y+EU1FMVFF7fHhtVUEmo4t9vxkrKdr6RvZOel3Fv6Wy9ASanpEKhS63LrIh6DsAAF7dkb6FVnDKCjRMu4jiGiZdastb1Ki2ophIVn11YkNNYaaOpb3jlOXq2nrVeel0Q2bMhX1lKTGapl16TlvVXtH8nKyutuJr0cXGriA3R/JK9R8yrGe//qJ4fHSUsak5XUP+1W+iJC+b2dVWFDe3ti0pyJU07jLBw2X8ZBzHi3Kz4l4+senZFwAcBwyx6dVPVOB9fLS+sRmVriFfCZf5qbdQKGMyVFdX11u3bk2aNAkAzpw58+7du1mzZkVHRz969CgiIqJfv34AYGZmduvWLWkW/g9a0GZmwGSKA4PRIsvYuDnL3Lx9ZgUCCAiAkBCIi4O4OAgJgZs3oa5Owb0rq2voNHGPtgaNxuPzOdxa2buUlFcEP3zSt4f97T9PGhvoTfdd274DllSvrKTTxe5DQ0ODx+NxOJwOW/tgs0rCJr2VzTZzaTSanh4DAG4FB3vN8j5y6MC/RKiRyyarU0RxFXWqUMDnNTRf2bqqCnUt3W6uHq7rjhKIZFFvBkmNoqahDQB5Mc+eHd/Sf3bHr44y4HFryGriGpHVqUIBny9Ro/pqlqqmrtWwSc4rDxGI5MhfNwNAQw2r6G24NtPWedUhNS1G5K9b2ytaU1VFoYoXdqHR6Hw+r5bLbcqlUKnaugwAePrg3uZl89f9sK9dxjk11WoUsXF1Gk3A59fXNhvHMAwjEF49CF75rUvc88ff+iwGADUKVUNbFwDePHlwZIPv3A2KdBKKXhIKpAehjCVVMAzz9/cPCgoKCgpiMBjdu3ePjo4GgAMHDmRlZcXGxsbGxsrwzvB/0YKWAZsN1RJPVRQKkMmK7vvwIdDp4OIi3jQwAAYD7t4FDw9pe5z48/L2A8cA4OD29dqaGmyO+PfE5nJJJCKNSpEtqKaqMnRAvwUzvgeA/VvWaNj2r6is0tXWUvSAJdDW1mazxQ/gbDabRCLROrRC0vGTp7bv3AUAB3/Zr62tJWGT08qmtFwWq3Lh4iW5eXlB1wP69un9xYVEqFDpvDrx/ZJXz8WIRLJqc1+Tfrce7vsDRPGBczf+Pc+lgV2lStdq4NS8OrubU1E80u+grqWdnHPXuZCpGrz6phrVYgQiSaJGutaO3+wWd3P1nrXupu/IBk4VgazC6Nbbcpg7APTwXHFjyfBGTrUKrR2NXA0trVqu+N7J5XKIJBKl5Wv26qrKH9etLCrIP/z7JbsevdpVI5qGZn2t2Hg9l0skktQord/hDx49scfAoQ+vX/ph0fQDVx8AAKe66tSudeVFBRuOnrPqLqd7QQwOuPQ5RXGhzPb1J/N/0IKWwYgRoKXVHK5elb9LE1euwKRJLVLc3WWP5fD1mV6R+LIi8aWPp7sV0ywlXbz6UUp6loWpidweBjNjw6Y4gUAgYAQikdiOA5bAyh8P5G0AACAASURBVMoqOTlFrJ6SYmFh0bH+jWVLl7BKi1mlxXO8Z1tZWianiEdHpKSmWjCZkjbbzG1sbBw9brydnW3Ei3DZTrPThETQ9U2qC7NE8erCbBrDWLIHrDw9sSQ5RhQnEskYgYgRSUI+L3Sfr6aJxfgf/vy3eWcAoOoZsz8MLGEXZVMZRpI1YmUmlaXGiuIEEhkjEAlEEkXHoHl/DMMIBIzQPndham6RmSa+Ftnpacam5pJXisdrXDJjsmXXbn8FP2yvdwYAA1NmfqZ44EdBVrqeiamk8UeBl+9fvQAANE0ttymzclLf8Xk8Po/3w6LpppbW+y4FK+qdAcRLXskI7Twt7eJrd9A4DiYmzZvz5kFoqDheVtZ6SKOXV3PJrCwYNqy1tbw8sLERx//4Aw60fFI+fBgCAhQ8rmEDnSoqq26HPq5vaDh05rzXpAmi9MC7oYUlpW3u4uYyODElLSD4PpvD3f7LsYF9emhpdLCXc/jwYRUVFbdu3a6vrz906JCXZMU7yvBhLhUVFbeCg+vr6w8dPuI1Y7oo/XrQjcKiojZzb9y6LRAI5s+bm5efn5Obm5ObK1SgMdIJQoZ2/RrY1XkxzwS8xqS7F7sMGStKz3kdVltZxuc1hB1eW/r+La+OG3vtlIFNLxUKLTfqCS4UdB0xiVtRwikv4pQXKXUlpPaiZ9O3kVtdFBcu4DWm3f/bfJB45EBB9OP6qnIBryHi+PqK9HheHTcp6LRu155kdZq+/YCa/Iz8N4/49bXvgk7rWDmQ29mr3m/QkOpK1tMH9xobGi6cPj5+sqco/dHdW2UlxY9D7ggFgskzvEsKC4ry84ry8xS5+k3YOw1iV1W+efKA19Bw6/xpl/HiJ9eIh3dYZSUa2jrXfztalJtVWVYaeMbfqnsPEpn8OixEKBSM8vAqLy4sK8wvK8yXMYT5A7icD1VwIWpBf4UQiYTAs/47Dp5gDnCjUykbl80Xpc9ZszUuKaXNXdRUVQPPHt17/IxZf9f4lPcXjravz66lOjEoKHDHjh3m5kw6nb5p02cYE0YkEoOuB+zY+aO5lTWdTt+0YZ0ofc68BbGxcW3mxsbGxr2Nt7S2aQqVlVX/BiGMQBjpdyAu4NeA5ePIalTHb+eI0p+f3snKTjW069tz0vyn/hsDlo+rLsweuvRHAKjITmHlvL++cmJTaOR+6iiOzwhGIAxatj/pxm/31n5LUqPYjPMWpUf9/kNVbqqeTR/biXMjTm2+t+5bdlFO/wU7AYBIVhm0fH/KnT/vrJlQnZ8uSmwXBCLx8LlLJw/8NNrJnkqjzV2+WpS+ffXSlMT4lIT41KSEcQN7NAV2tfyr32ycQNxw9Nw/J35Z6NZPnUqdPF88QOv41tVZyYlOI0aPdJ+6zWfyim+H5rxPXv3zSQDITE7ITklaMmZAU+DIHTeCY00Le7eJuG2nNDClzvSBwMuy8fpPff+mEAxmJ63qLeB10qKxnUJlZZWugVH/WWvsxkxXtlanrertNNbNyNpK2SqCxoYbS4at2LxjztJ2D/BoLyWcBk5jJ63qPc7WQJ0s7jys5bC/HzPy5pEd0grnFZftPBd4IfC2kg7m//slIQKBQMhF9oT9QiXOw4McNAKBQEhD1Ikhc8J+6eOgPx3koBEIBEIaGC4axSENXKhM/4wcNAKBQEgFBxxkDPbAhUJchvv+ZJCDRiAQCGlgclb1FgoxTIlj4ZCDRiAQCGng8ufiUOagJuSgEQgEQgo4BgAyvjnCcRx1cSAQCMQXQdSClvGSUIAROjjjgiIgB41AIBAykDPMDrWgEQgE4gshbxQH6oNGIBCILwSOg0B6G1kgBNTF8d8FxzAgq3aCUANO5Dd0xmQF6kRMqeOKmsAAx5X6DQAAAAgxAgCQSQSKihL/zURoG+ipayiwhMcn01BXlxYdp2wV0cqzfAFez1f6D6+U08iq/TxLG8sGB+AJcPXmaeFFzWMZbWQcBG2v+PxZQA5auWBkVfzTlpJSkFq+oLFT5iYkqZPJhM6oUedMyIQTiACgQiRQVJT+v9Ctdw9ep1yjN/dCi9KVPyuTUAAAfCFez1P6xKolNfX51fXKVhEhaNFlgQHIGqeB4wKlNiOQg1YyndLYBIBOaW4CdJbf/CpRUSHzOmtKtq+OTvlxt6kkbxz0Jy7mKRvkoBEIBEIKovEbsmazA1xGD/Ungxw0AoFASAEjAI7jAqkTiuICfntXAmsXyEEjEAiEdGR3ceBCBdbN6jjIQSMQCIQ0RIM4ZE7Yr8zFJ5GDRiAQCBnIXDNFKMCIShygiRaNRSAQCCngAE0rw0oJOF96DzWOr1ixws3NbcKECRUVFa1ya2pqmEymbH3koBEIBEIWuFAoI8jYMSwsrLS0NDQ01MPD4+DBg61yt23b9rHXbgVy0AgEAiEd0UtCGUH6p97h4eEDBgwAgIEDB4aHh0tmRUVFsdlsMzMz2eLIQSMQCIRMpHvn+kbe4zdv3dzcNmzY8PF+LBZL1InBZDJZLFZTukAg2Lx5888//yxX+V/moDEMCgqaN3//HdzcxHE9PcCwFiEjA0aMaJ0oaWHQIEhNbS1hawt6evBxtxGX22xBRQWcnCAyUupxvn0LvXsDANy5A5MmfUKFEQjEvxdcNCO/dFSJhBF97ENDQyW97Z9//jllypQ9e/bo6Ojk5eUBQG5uro6OTlOB48ePe3p66unpyT2Af5mDls3jx1BV1RwsLSE4WBw3N2+OGxtDWBjMmwcREa0txMZCTQ1oaUFYWNsSOTlQVQXZ2fDNNzB1qrIrhEAg/t1ggMtqQYNQCB996u3j4xMQELBly5YhQ4ZERUUBQHR0tLOzc1OBmJiYgICAsWPH5uXljR8/Xob8f8pB0+mgqdkcCASgUpvjNJo4jmEQGwsUClCprS1cvgwzZ8K0aXDlStsSIgvGxuDnB7m5UFsLkZEwaJA4VzLeiuRkMDODV686VrOY+ESnkeMZ1j3mLl/b2Nj25FjOY9xTP2EGnIS3caNdBtkxjVYvXchrbHtisImjhmWkve+A8diYmP5OTvoMxry5cxs/Mt5m7u3btx3s7bU0NNxcXVNTUpQnJGKIs/P7jx+npFOSkXzRz/OEl3OI/1YBv+0rwirI9p/av2kz/v61MwtG+3s6Xd0yl1WQraBQVU7qox+8by93izr3o/AjoeeHVt5Y7CIKL/3XihLrq8ufH151a5nr4z3zOSW5ildKNvy0YLyh+tPtpCXFL3Qf9W2/bj9vWMHntb5GLx+FeI92HtvDwm/W5NzMNFHiujmeox3MRGHLolkKCpVmJl9eM/XXWUNCj22Tdo0qC7JPTh/QtJlwP+CPxWN/nTUk5PBGXn2dXAlMNIxDNtJHcbi6umpra0+aNOnatWtr16599+5d3759AeD8+fP379+/d++emZnZnTt3ZBzAf8pBK86aNXDsGGhrt0jEcfjnH5g1C6ZNg6AgkOKkAAC4XLhwAYYOBQpFIbmCAhg/Hk6elOq+ZSIQCNy95vnO90548bCgqHjvkeOtCjwOfzl/5bqIqNgOGG+S8J7mMXfhkieRsUWFhf4H97cq8PzZE79li6PfvO6Y8Unu7r6+vm8TEgoLCvbt3Ss3t6SkxGf27F8OHMgrKHAeMmTatGlKEgKAx48fL5g/P/Ljxynp4ELhjT3Leo2f4XP8BqeiNDLgTBtlcOF9/22CDw6oqjgv7LefRi/7Yf5vIQym9aNfdyso9PLY2i4jp7jtvlxXWZZy53yrAuziHOfVR1x3/uW686/es9aLEl8dW29gP2D0TwG61o4xF+T3Y8o/DE6RIPc5zi37dFNCgWDzolmTZs374154eUnRxVNHJXMry8v2rPVdumnX9VeJjn0H7FoxX5Sel5X+8+9Xzt5+fPb241W7FKoRLhTe/ml5z3HTZ/kHcSpK31w720YZXBh6fHvTNcpPfPM64LdRvru8DgWQ1Sgv/joiXwUwHMeFfIHUIBCC9HHQGIb5+/sHBQUFBQUxGIzu3btHR0dLFkhOTpZ9AP8+B929O2hri8Py5S2yhg9vzjI1bbflFy9AVxccHcHeHkxM4P79NsowmWL769aBAl34AACVlTBmDLi4wMSJ7T4kAAB49jJSU4PuM8PTyEB/y5rlV67falUgNiGJoq5OVfBu0RavXoRraGhOnTnbwNBw1bqNQdeutiqQ+PatujqFQvnomUMBnj17pqGp6e3jY2RktHnLlistn07azI2MjOzRs+fYsWPpdPqWrVvfJSVVVVUpQwgA4mJjKRQK9ePHKenkJb5RpdAdXN2p2noDPRelPLv7cZnY4L/pDIOmTbKqOlFFVYVCIaupk1TU1GiaigiVp8aQ1WkWQyaoaTLsJs7Ji3wgmYsLBQ01LEbXXnRDJt2Qqa6tDwDV+em8+tqu30xX1dBx8Fjac7qf4vWSBl5bAQQSED7DZ2tvX7+k0jXGekzX1TOY5ev3KDhQMvddXLS1rf2A4aMoVNqsZX5ZaamcmmqBgF9ZXtaj30AzS2szS2s9Q2NFhPKTolQo9O6u7lRtvf5TFr4Pb+Mavb1zmabbfI0KkqK6DBhp5tifrmc0cOqS98/vyZfBccABhLjM8H/1JeH9+2BkJI5fuQIPHzZn/f039Oghjndgir/LlyElRWy8qgr++acNlxoeDhoaIBDAs2cwZgy8btmibPOLz6dPYc0a+O03+OknMFbot9WKjOwch+62oriDrU1Wbi6O45JzGPotXQAAN+8+aHt/BcjJyrTtbi+K23a3z8vJbiWxePlKAAgJbn1vUITMjAxHBwdR3N7BITsrS9J4m7kjR44c9OFpIzIigmlhoaWlpQwhDMNW+/kBwK2bNxWvUVVxHoPZVRRnMK2rSwoAxyV/ctXF+W9Drk7efirtlfj3SdVmuHivvrR2BoYR1Oiac0/dVkSIU1agYdpFFNcw6VJbXigpVFtRTCSrvjqxoaYwU8fS3nHKcnVtveq8dLohM+bCvrKUGE3TLj2nrVK8XtIg6DsAAF79GXpLCnKzrWzsRHHLbrbFeS1+zH0GD7Xv4ySKv4uLNjQxo2loFuXlqKqpb/Odk52WYtejz+KNOxgGRm1bl6Ba4hrptnmNSvLjQ666bzuZESG+RtomlmkvT1eX5FM0dZMeBdXVVPEb6kmqajJUROZw6Z96i94hyj3aDvPva0GbmQGTKQ4MRossY+PmLHPz9pkVCCAgAEJCIC4O4uIgJARu3oS6jzqhzM2ByQQrK/Dxgb594fFjAGge8pGf34bl0aPhwAH4/nvYurV9h/SByqpqOk3cvtOg03g8Podb2zFT0qiurKLR6aI4jU7n8XhcLudzGa+srGwyrqGhwePxOByO7FwajSZ6hX379u1ZM2cePiL/YbNjQh2rUQOnRkVdfEVU1GlCAb+xXuKK4PiDEzuHz12nIvHAwSrIjrh6esYvl1Zcfe3oNjnk6DZFhHjcGrKa+MGIrE4VCvj8hubfZH01S1VT12rYJOeVhwhEcuSvmwGgoYZV9DZcm2nrvOqQmhYj8tcO/uqUBKemmkKlieIUGp3P59XVcpty1SlULR1dAHj5KGT36sXLt+0BAFZZqQ5D/9tps/ed/ZukQt75od9DNg2cGhV18alr8xo9OrHTZc5ayWvUzXm0sV3fC77fnp07soFTAwD8RjmLAOCY3C4OAUZUYjP339eClgGbDdUSLzEoFCCTpZduycOHQKeDi4t408AAGAy4exc8POD6dRg0CDQ1AUBsXyCA8HCIiICdO0FLC96+hYQEMDODEyfasKyuDgCwezfY2sLy5eLhd/I4cfb89r0HAODgj9u0tTTZHPGPmM3hkkgkGrXjvRltoqmtxWGzRXEuh0Mikagf/os+HW1t7SbjbDabRCLRaDS5uSwWa/HChbl5edcDA/v07as8oQ6gStNorBNfkcY6LoFIVFFrviLxodepOnqWfYfWsZu7ZTJfP7HsM9SoWw8AGDzD9/j0QY21HBWKnAMgUzV4H9wKr74WIxBJqupNubrWjt/sFnfU9J617qbvyAZOFYGswujW23KYOwD08FxxY8nwRk61imI9Kp0ATUOz9sO9v5bDIRJJ6i37zdjVlb9s9istKth96nw3h54AYN/H6c+Q56Lc1Tv3j+ttVV3J0tTWAZmo0jQa6z6cuo+uUeLDQKqOvkXfofUS1wgwbOTircMXbMRxvK66Mu7O3wr0RH34pFtqPv7xq93PyL+vBS2DESNAS6s5XG3dkSqLK1daD1h2dxeP5fDxgdgP79+YTNDSAl1dWLoUdu2CoUPBxgZ8fcHZGYYPh6VLpdo3NoY1a8BP0Q5B3/neFRkJFRkJPjM8rZjmKe/TRekpaRkWZqaffY0GpoVl2ocxDOnvU03NmZ9RwtLKKvnDMIzUlBSmhYWk8TZzGxsbx44ZY2tn9/LVKwW9c8eEOlYjLQMzVr54wAwrP0tD30Ty2Tkv/nXG6ycnZ7mcWzxBKBCcnOVS9D5eKBTiTbOa4bhQwJfxXNwEVc+YXZgtirOLsqkMI0khVmZSWar4l0kgkTECkUAkUXSaO1UBwzACQanzEbcXYzOLnAzxQKDczHRDUzPJq8DnNa718WRadzt1LUTknQEg+W3M29cvRXGSCplAIJJI8huOmgamzdeo4ONrFJn55slv3i7nl04UCgS/ebsUv08oz37/8qI/gUgiksiFyTEMi27ye0pFk9nJ+tQbV+pqL/+iSwsAgONgYtK8OW8ehIaK42Vlracp8fJqLpmVBcOGtbaWlwc2NuL4H3/AgQMtcg8fhoAAAAA2G8aPByq1hfGCAli7trlkTQ3ExcGUKeKBdD17in36+PEQFCQutmOHuEuknQxzHljBqrwdElrf0HDo5G9eU8Q3ksDb9wqLSzpg8GMGDXGprKx4cDe4ob7+12NHPKZOF6XfuXWjpKjoE40PGzaMVVFx+/bt+vr6w4cOeX24LoGBgYWFhW3m3rx5UyAQzJs/Py8vLycnJycnR6hAR14HhDqGmaNTHbsq4/VjfmND1M3z3YdPEKWnvQzlsEpdF2+Zc/zm7CPXpv70B0YgzD5yTd/KzqqfS3rEo/x3MY21nPC/jhrb9Val0uUK6dn0beRWF8WFC3iNaff/Nh80RpReEP24vqpcwGuIOL6+Ij2eV8dNCjqt27UnWZ2mbz+gJj8j/80jfn3tu6DTOlYOZIp8oU6j14DBNZWVLx+FNDY0XP1fe/cZ10TWNQD8TAIIpNCrIoIIKhYQFBSxt7UrVhTBxY4Vxe7a0BXX3te14drrWlBUdO0dpSNFlN57DSQz74cJESEJQQ1P3Pf8f/NhvHe4JzfgyWTmzr1HD/QfMYYuf3znZm5WxtN7t0mBYOg4t6z0tMzUlMzUFIokeRXla+Z4RAS/LispPrZzS3t7Bxan/qV1m7XrXFFckPD6oaCK9+6af+uewtHE8S+CSvOye89c5bbnH9cdl1x8jxEMhuuOS3rmrdm6BqG3zsa/DCrKTH11/lD7AWPq7w8BQAEpoCRtlICS66reCpag/19iMplX/v5r7e/bTTs4ctjs5Qu96PKpcxeHhEf+qBDHz1zcummjXVsLNocz39uHLl84a3p42Pcu/8xkMi9fubJu7VozU1M2h7Ns+XK63HPq1JCQELG179+9Cw0JsTA3F235+fnyCPRtCAZj5Mo9z87sP+zZX0VNvcsY4VXR27tXZyV8UGVrsHX02Tr6LG09giDYOvpMJWVdU4v+c9be27f2z1/75aclDvbeImOgrnO3Rv5z+PaS4Uqq6laD3enyt0c3FCTF6Fl1aj3s15cHV972GV6cnthl+joAYCqrdJ239UPAiYDFQwtT4ulCxcFgMn0P+h/b5TfOuaM6iz1p1gK6fMvSeXFR4bGRYfHRERN6dRJtxUWFNg5Obl7e6+ZPG+tsk5QQt2r7AVkCEQzG0OW7X5zdf3RafxU1lr2L8Hd0d8/qrE/Rot8Ru8bvSJWtMWC+76MjfhdXebRyGthOlgRNAQVS5+KgSJDnJQ5Cli9i6JtRRVlUVWOsRpyvpFlJqDRCIA01ZRVm463gKW/5+fn6urq9py3rNGyyvGNlF1WUNsqisY22qndV2MkZPmsmzphX/8Hf521yYUph/c+V/BDrBrbWVBPe3CotzB/h2On0rxJH0KYWFP/x7MPZh9/4hFq9fqqbhAgh1KgIoChSIPH6m5SqHwITNEIIiUfVv+QVSFlS9vthgkYIIfEIICiKoiSfJlMkieOgEULof4ECoEDKs4IUSVJ8Od5XwASNEEKSUAAURUp+1JukKAoTNEIINT6CAAoogcQUTAkEhFIT+cXHBI0QQhJQFAVSH/UGoKp48ouPCRohhCSjgBRIvsQhkO+j3pigEUJIAoIQruotCSltwv7vhwkaIYTEox+0ljKKg6RIqs6yXj8QJmiEEBKPHgctZSYvHAf9cxMwVEiVxnmTlZQbadpJuY77rIFBgOThTT+KQEABwJuLhyNun5V7LJJqnElMeBW8JtAYc+xUAfjv33nu2GF5B2IoKTXWrKqE4769E0YPF/6LIoECqcPspJ1ffz9M0PLFb8KWfIPhR2I2ThgAgYASNMp/fmiUjwF1DnfKzLlpKcmNEIsCYPzomb7F4lXxGfJPZ6RA8PhuQKWAUSVQq//o72NqaqKhIfdZVSmKCn14OzIsBKoTNEX/vqScKJCUlEF43w8TtJzRNxkaIQ40TtastfDbT48giBUbNpXwGuPTgMEgGucksLRSUCXnSXwAgFdR8bh9c4aOJUO/vbxj9XAbaWPfTt5RSIFgca+vVpIlACiKIiWnYFIgIGRf16nhMEEjhJBkUpeFpSiSqsIlrxBCqPHRX0tJSsomfc3v+fPn9+/ff+jQobm5uTWrtm3b1rt3b0dHxxSxS1FXwwSNEELiUcJh0BJRJEUoSVwo48GDB1lZWffu3XNxcdm+fbuoPCws7Nq1aw8ePFi5cuW+ffukvAC8xIEQQpLIMFlSpcQlk548eeLg4AAAjo6Ox44dE5XfunWrd+/eBEEMGzase/fuUsLjGTRCCIlH0KfQAlLSVsXnx2bl+vn5nT0rZphmXl6eqakpAJiamubl5YnKMzIyPn78OHDgwEGDBiUmJkp5AXgGjRBC4lH0qt6Sz6AZFGiz2XZ2dnp6eqLCEydOBAQE2NjYaGtrJycnA0BSUpK2trboABaLVV5efvLkyXfv3k2bNi04OFhS+5igEUJIAvoGoJRx0BToqKr069evZpmHh4eHhwcABAUF+fv7A0BwcLCTk5PoAAcHh2fPnjGZTC0tLULqqFW8xIEQQpIQFEWRAlLSJv1R7759+2ppaY0aNerSpUtLliyJioqys7MDgKFDhxYWFjo7O7u6uu7Zs0dKeDyDRgghCejpoKU/6i15mB1BEDXzr66uLn01g8FgHDp0SJb4/60z6FOnwMkJOBywsIBdu4SFoaFga9uwdpSVoUGPb3btCjExDQuBEPoZUMIJN8RvQALFx1W9ZbFzJ+zaBQcOgK0thIeDpydwOODpKd+gDx7A6dPw8qV8oyCE/icIAqSv6i0gGSpyXPLqv3IGXVgIGzbA1aswZAgYG8PAgbB9O5w//9Uxr15B166190kSvLxAWxv09MDXFwBg4EDg86FlSygrgydPwNYWdHVh8mTIyAAAiImB3r1h0ybo2BEA4P17UFcHFqtBLzbk/fvujl2aGerPmu5ZWVl7MlmxtSOGDNbV4NDbuNGjZAwUGvK+VzcHM2MDr5nT6gaqW3tw3x5DbW7N7fXLFzIGcu7q0NzIYPYM8YHq1o4cNkRfi0tv48c0oEcNDZSRkT56xLCmBrp9ezrHx8XJGOhDRKjb4N5925uvXzy3qs6Ev4/v3R7b26FH62azJ4z4/PGrNn8dOSDxY7yMUWqJDg91/aVXT2uztd5edYM+unt7VK8u3Sybzhg3/HO8rB0RiYsMmzGy33B7S79l8/l1Gn9+P9B9oNMvHVp4u41OShA27jN13MB2JvS2aqbbt3WqLn7cTYpX+KNao6XERm73HL5qcKezm5cKJDx7nZWUsKx/w6YNoapns5O4URRZ5+/wB/qvJOg3b8DYGDp1+lIyfjzcvVv/D169Cvfvw/v3cO8ebNoECQlw5w4oKcHHj1BRAaNHw8aNEB8PHA5MmSL8kZAQSE2FU6cAABYvhr17QUtL9lcqEAjGjR41c7bX25DQtNS0bX5bZKmNi429euPmizfBL94E79y7V8ZAE8eMmj5rzst3IelpaTv+8Ku3dsKkyU9evqW385f/MTMz79DRRpZA411GzZw95/X7kLS0tO1bawcSWxsXG3P52o1nr94+e/V2x25pD1N9Z6CJY8f06dv3fXikg2PXhfO8ZAlECgTev7qO85h2/v6L7Iy04/t21qzNy8n+bcGshWt8A99+sOnsuGL2VLr8zfPHG33mhb97K0sIsUEXTp043mP65X9fZmWkH927o2ZtbnbWqgUzF/+26d67GJsujktneTS08ZUz3Ua5eR6//SQnM/3Uwd01a/Nzsjct8ZqzYv3lFxHt7RzWz59Glyd/ivc7eu7IjX+P3Ph34Xo/cQ03DFWSLkh6SpVmf39TNZGk4Mjy6c4uU5b9HViYk3nv7wNiQpPkud+X8Ru8fiBBgdQELfVR7+/3X0nQSUlgYvItP8jnA0VBURHY2EBKyleN3LwJvXvD0KGgqQk7dsCzZ8KVbwQC2L0b2n/jDF5PnzzmanDd3N0NDY2WrVh54fy5emv5fH5WVmY3p+6tLC1bWVo2bdpMlkDPnjzmamhMmuJuYGi0ZNmKyxfO1VurpaVNh2hlaXnuzKl9f/6lqlb/TJJPnzzmamhMnuJuaGi0dPmKixfq9qh2LZ/Pz87KqtGjprL06BsCRYSHl5QUz52/UF/fYN1GX79tOyS0/ZXgV8/YHO6wcZN09Q085y+5c+1Szdrwd28s27Zz6tNfnc32XLAkIfZDlojP9QAAIABJREFUcVEhAMREhKuqqaupq8sSoq63L5+xudwR4yfp6htMX7Ak8J/LXwd9a9W2Xfc+/Vls9oyFPh+rg8oo9PVzFof7i8tEHT0DNy/v+zev1KyNCgm2aG3t0KufOovtNtf7U1xMSVGhQMDPz8nuYO9oYmZhYmahZ2j8bf2qiSrLBYYSMH7wldWPIa/VWJwug8dwdfT7u3u9C7pR95gnl09q6Bs1tGWCIICiSIG0jcBLHPUzNBReghCpqAB/f+BJ+MAUfeiNGQOzZ8Pw4dCyJfj7fzU1aHIyBASAkREYGYG5OTCZkJUljPUdEwwmfEywbidM7m2trRM/f675CSy2Njk5WVVNzXXc2PZtrKZOcUtLS5Ul0KdPCdbWwhka29QJJL329csXAoHAzr6zTIESvjRVt0dia1OSk1VV1SZNGNvRuvWv7m5paWlyChQZEd6qleWCuXM6dbD2dHfT0NSUJVDq588WrdvS+y2t2qQlJdUM1Nmpp9+f/vR++Lu3Rs2ac7gaADB5hpfPBj+Ohkwh6kpJ/NSqtbUoaGpSYs2gXbr33Hb4JL0fFvzWuDqojFKTPptbtaH3zSxbZyR/1aNO3ZzX7z9O70eFBBs2NWFzNbLSUpuoqq3xmjqpbxffRbNyMtO/rV81MfTbMZs5EswfnNFyUhONWlrR+0ZmlnnpybXOanPTkp9dOzN0lk+Dm6Yo+mFvKRtZIfFR7+/3X0nQdnYQHw8REV9KgoJgxQpo8vWfguh+q2gGqeRkcHGBT5/gyhU4dQquXv1ysKEhjB8P6emQng5paRAeDoaGAPCda0QWFORzOGx6n8PlVlVVlZaUSK/NysgwMDD4dfr0K9duqKiouLlOlClQfj6bLZzjnMOpE0hyLUVRK3wWL12xWvYesTmSA4mrzczMMDA0+NVz+qWr11VUVNwnydajhgfKysq8fSvAxrbTpavXDY2NPSa7yhKoqLCAxRb+FlhsDp9fVV5aKqpVZ7G0dHQB4PG926vnTV+8/ndZ2qw/aEGBOksYlM3m8PlVZeKCPrp7e+XcaT4btohvRYKSokJR4+p0j8q+NK6mztLU1gGA5/cDfRfNmrdmEwDkZWdp6+oPnzBly5EzSirK66qveyig8uKiJmrC3qmyOAI+v7K8TFRLUdR5vxUjvVaqqn/LlP8UgLQzaJKS61Iv/5VRHAYGsHAhjBgBe/aArS3ExMDChTB37lfHaGpCaCiEh4OJCezfLyy8dAlOn4aAANDSAiYTRP/hi4th8GBYuRIePQIbG9i1CwID4YVMd8yk09TUKi4WRikpLlZSUhLlAkm1Dl27BoeG04U79+w10tXOzcnR0dWtJ5CWVklJsbCpkjqBJNcGv3nN5nAsWrWSvUclxZIDiat1cOz65n2YsEe79xrr6+Tm5ujo1Nejhgdq0kTVqbvzVM9pAOC7eYuRrlZeXq62to70QFwNTVHqLy0tYSopqX19H7ioIN932cKM1JRtR061aV//ZXpZcDU1y0q/Cqr+ddDCgvyNPgvSU1N2Hj3dpkPDgrK5GqLGy0pKmEwlNfWvGi8uzP9jpXdWeqrvQX/Ldh0BwLpT5xOBT+naReu2DrY1L8zP09DSBsWjxuHyyoW9qygrYTCZKmpfLjS9vHleQ9egTddepYUFDW6anrFf2igOChhyXNX7v3IGDQC+vuDtDWvXQqtWMH06eHrCsmVfHWBlBV5e4OQEvXrBnDnCwpkzoXlzsLICW1uwtRXeCRw7FkxNQUMD/P1hzhwwNobHj+H06R/yMs3MzWI+RNP7MTExzU1Naz7rKbb2zevXTx8/pgtVVFSYTKayDNdYWrQwi/nwgd6PqxNISu2Na/+MHS/TKa2wKTOzmBhhU7F1A4mrffvm9dMnwh4p0z1SkqFHDQ/UrNmXmwoMBoPBYDBl+ALU1NT0c3wsvZ8YH2fcrHnNQFVVlXMnuZhZWJ64fu9HZWcAaNa8RUKccDT9Z3FBZ7uONmtl+ffNoIZmZwAwNmmR+FHYo6SEeMNmJjUb51dVLvEYZ2phefBSIJ2dASA69F3o6+f0vpKKMoPBVFJS0PM5HePmmZ8/0vtZiQnaRs1q9i4u+EXEs6DVQ+03u/YlBYLVQ+0To0Jkbbq+URxAUmRlQ288NsB/KEEDgJcXvH0LpaUQHw8rVgivRXTsCO/fCw/YuROKiiAkBMaOFZ4Oczhw7RoUF0NuLhw+LLy4fOYMFBaCujoMGgSRkVBaCvfvg7k5AICVFURH146bnAxWVjK+RucePfNy8wJu3qioqNiza+cE10l0+bWrV9LT08TWVlSUTxg35uXz58VFRRvXre3azYmrUf/1x+49eubl5d4OuMmrqNi3e+e4CcJv99f/uZqRni6pFgBuXLvap19/GbtT3aPcWzdvVlRU7N21c/xEYVPX/rmanp4utra8vNx13NiXL54XFxX5rl8nY4++IVCffv0iIyKuXL5UUlzsu35d5y4OGjJcI7Zz7F6Qn/f43u1KHu/UX/t+GTWWLn9w60Z2ZsajO7cEpGCk65TM9NT01OT01GQpqz7Lzr5r98L8vEd3b1fyeCf/3Ddk9Di6/P6t69mZGf8GBpACwWhX98y01PSU5PSUhgW1cehWlJ///H5gJY934eiB/iPG0OWP79zMzcp4eu82KRAMHeeWlZ6WmZqSmZpCkSSvonzNHI+I4NdlJcXHdm5pb+/A4nC/v5vyYGHrUFqUH/E0iF/Je3juiP2AkXR56KPAwpzMMd7rl/991+d4wNy9ZxgMps/xgGatrGVtmmDUO4qDIXk+6O/330rQPwMmk3nu0mXf9eusWppxOOwlS4Wn+TOmeYaGhIitde7Rc/nKVW6uEy1bmsXGxBz195cx0OkLlzdvWGfdypzN4SzyEQaaM8MzLPS9pNrIiPCKigpj2YZViAKdvXh508Z1bSzM2RzO4uqmZk33DAt5L7bWuUfPZStXuk+a2LqVeWzMh7+On5BTIFVV1bMXL23fusWyZYuIiPC/jsv01jGYzO1HTh3a/vsQh3bqLI7H3EV0+frFc2Iiwz6Eh8ZGhg/v2lG0FX/Dd2dxQXceO31g2+aBna1ZbPav84RBf1s050NE2IfwsJjI8MGOHURbg4IymEzfg/7HdvmNc+6ozmJPmrWALt+ydF5cVHhsZFh8dMSEXp1EW3FRoY2Dk5uX97r508Y62yQlxK3aLmbsmoJgMJiem/+8fXTXutFOTdRZ/SbPpsvPbvJJjYtS52pq6Blo6BlwdQyAAA09A6bsN/lJqt5x0BQpxwUtCbkO4kMVfIFAylRYPw5f0Ei/SAb8pxaNBYDUonJcNPYb8CoqBrVvzjS2b4RFY13dG23RWMvVa9dvXPcbXVKclzfQwuIPa4ln3Jk83onCwutym+lBQS8qIYTQ/xwBQFFA8iV+2pF8kqFa/7MC3wwTNEIIiUd/K5W2LCzIdxw0JmiEEJKAooCqb01CecbHBI0QQhIQBP2giqR6SgAMFTmO4sAEjRBCEtQ/YT9F8vASB0II/U9QlNQ1CSmQ5/M7mKARQkgCggAKKGmXOCiCIcdhp5igEUJIAoqipI/ioCiyUvz6AD8EJmiEEJKEAApIvsQETQooQhlvEiKEUKOjSPppbmln0BQfz6ARQqjR0bPi1TMOWp5zcWCCli+KAqJRpq4orKgqq5Lj8u8i3CZK1Q9YyZcSg5B8b+ZHKq8S8BtnvhQ+WfUj5r2rV1YRr4gn9z+GKl4FAJhbmJo5dJN3rPIK3r/3nss7Cp1q+YIaCZde8kryo94Un2Jw8FHvnxa9plkjBMoq5RWWy/GrlkglW0VZnretRZgMonHmf6oUUI0wtRAAFPL4PMn/1X+guNyynBI5TlJM41fyAMCqbcu+w/vIO9YZ/39C3kbUf9x3okgA4Nc80aEvb0j5O6RAUF4usZKiFixYEB0d3aRJE39/fx0d4UoRfD7fw8Pj48ePysrKJ0+ebNGihaQWcLpR+Wq0ed8aLRDZKCebAI3zuQbQiG9do3zxoAPhFJXfrubfA/0+kgJS4kaSUmZ3fPDgQVZW1r1791xcXLZv3y4qv3nzprKy8osXL2bOnOnnJ225dEzQCCEkHv0NmBJI2xiSR3E8efLEwcEBABwdHZ88eSIq53K5RUVFJEkWFBRwudKWQcAEjRBC4lEkSY+DloRHks+TEuzt7ZcsWVL3x/Py8kxNTQHA1NQ0Ly9PVN6rV6/09HRLS8uVK1fOnz9fygvAa9AIISQeQRAUBVJuVisB4Wjc/NbbtzULT5w4ERAQYGNjo62tnZycDABJSUna2l/W2922bVv//v3Xr1///Plzd3f3oKAgye0jhBASiyAAQMpqRQKgqDrr5Hh4eHh4eABAUFCQv78/AAQHBzs5OYkOyMnJoW8M6unp5ebmSomPlzgQQkg8erpn6ZuUR7379u2rpaU1atSoS5cuLVmyJCoqys7ODgB8fHyuX7/u7Ozs6uq6d+9eKS8Az6ARQkg8gkFQIO0Sh4AChmoTiT9OEHv27BH9U1dXNzg4GAD09PQCAwNleQGYoBFCSDyKIgFAyth16WfQ3w8TNEIIiUdRQAFIWdaKpCi5PhmACRohhMQjGAyKAsmT2QGfAqaaqvxewM9zk/DcOejeHTQ0wNoaFi+GsjJhuZ4eEMRX28ePAAD+/mBuDlwujB4NovukYgtppaVfWlBRgc6d4dUriS8mNBRsbQEAAgJg1Kgf31mEkCIgSRCuHCtx45fLccmrnyRB+/qCtzfMnQvh4fDXXxAZCV27QlX1pZ9//4WCgi+bmRkkJMDs2XD+PEREQHExrF0LAOILa0lMhIIC+PwZBgyA8eMbr4MIIcVDCRdUkbiRFAV1htn9QD9Dgk5MhN9/hzt3YMIEaN4cunWDW7eAyYR9+4QHcDigofFlYzDg2TOwt4fOnaF5c5g+HcLCAEB8YS10C8bG4O0NSUlQVgavXkHXrsLamvu1REeDiQm8eCFLh0Lev+/m0NnYQG/mNM/KykpZak/9fbKtZSsDHa0JY8fkSR04WVNcZNjsUf1Gdbb8Y/l8flXtQM/vB04d5DS0Y4slU0YnJ8TRhXnZmSumTRjeqeW8cb+kfP4oY6Co8NAx/Xs4WpmuWjCnqk4g2sQh/T59FEY5efiAralBze39G8nfV2oGCgsd3d+5i1XzFQtmSwo0fnDfT/Fxon9mZ2ZMnzja3qLp+MF9P3+Ml7FHsZFh04b3HdKp1e9L59UN9CwocPKAbgPamy6YNCqpulOLPcb1a9uM3lbMmCxjoITo8KXjB7l3b7t/zSJ+Ve3bTQ+vX5zzS1e3rlZ/LJpWXJBPF755eHfBiJ6THFqtmzY29ZOsPcpKiD67ePwht+739q4RSJjCOD/184GJDqJ/ht+5eHzWL4fcugfuXF5VIXFKIClSYiO3ew5fNbjT2c1LBXV6J3xhSQnL+rf/hsbrxY+7SfEKv6eF6gdVpORoYKgo/6gXXNfPkKAfPoQuXaB9jV8hgwGzZ8OdOxJ/ZPJkePgQKApiY+HOHWFWFVsoVmkpnDwJzs6gri7TK0xNhSFD4MABaW1WEwgEY0aNnDXH611IWFpa6h9+W+qt/fQpYcFcr7/PnH0bElpSUrxxw3pZXhQpEKyZ5TZisufRW09yMtPPHNpdszY/J3uLj9esFesvPI9oZ+ewYf40uvy32e52Tj397760tu2y6zcfWQIJBAKvKRMmec648fhVZkba4d3bax3w6unj1Yvmhga/EZWMGDfxyoOn9Hbw1AWTFmZt2neQJdDsKeMn/zoz4PHrzPS0P3fVDvTy6aNVi7xqBgIAL/eJ3Xr2ufPifafODmt9FsrSI1IgWDF98ugp004GPs3OzDh1YFfN2vycbN/Fc+au3PDPy8gO9g6/zfOky5MT4v84fv7YzYfHbj703rBVpkCkYMs8j18mTt119WFeVsaVI3tq1mamJB7euNz7j4M7rzwoLy05f2AbABTkZu9ZOd99ydojD963tu2yfclMWQJRJHlj87yOgye67blakpv15tIRMcdQ5L19vwmqP41SIt68vni4n9f6STsuKquqP/t7V90fqbd3R5ZPd3aZsuzvwMKczHt/HxD7ws79voxf9YPn3qNK0gVJT6nS7O9thwQAIKVuVf/fL3EkJIC5ee3Cli0hvvrcoVcv0NISbs2aAQAQBDAYcPEitG4NgYFAPyYvtrAWU1NhOz4+IHWWqS/y82HQIOjRA4YNk+XwJ48fa2hoTHH3MDQyWrZi1YVz5+qtffH8eSc7Ozt7exOT5lM9p0WEizv3ryP0zXMWhzvIZaK2nsGkOd4PblypWRsdEtyytbVDz37qLPZkL+/P8TElRYUJMVFlpSVjps7W0tWbtmSV1+pNsgR6++IZm8MdNWGynoHhrEU+AVcv1TogKjxUTU1NrcannYamllnLVvR27eLZTbv2q6rWP6numxdPORzu6ImT9QwMZy9aevPqxVoHRIeHqX4dKCYqorSkZOqsuTp6+t6r163aJNPvNOT1cxaHO3jMRB19A3cv76Cv37rIkLct21g79uqnzmK7z1v8OS6mpKhQIODn5WR1tHdsbm7R3NxCz9BYlkCRb1+qc7i9R47X0tN3mbHg6a1/atZ+eP+mpXVHi3Y2ukZN+7lMSoyNBoC4sHctLNt2cu6jxmKPnbko+WNsaXFRvYFSIt+qqHPa9h3J0tLrMnZG7JNbdY8JDTjL1jEQ/TM18m1Lhz4m7btw9Iwcx8+OfXpblh7V9DHktRqL02XwGK6Ofn93r3dBN+oe8+TySQ19o4a2XC+qLBcYSsD43kEQBEFRACQlcaMoYCj/Pz+DNjGBlJTahamp0LSpcP/MGQgJEW7Pa8zqPW4c5OTA/PkwYEA9hSJPnkBICHz4AEeOwKBBEBv7Va3Y0TaPHsHAgXDlCqSlydKbTwkfrdu1o/etra0/f/5Uc0EdsbUTXSfdCXpAUVRcXGzQ3bsOjvWfpwNAetJnM8s29L6ZZeuMlKSagWy7Oa/bd5zejw4JNmxqwuZqJHyIMjFruXPNEo8BXTd5z2JzpM2zJZL0+ZNlm7b0fqvWbVOSEmstETR19rxVm//gamjW/dn3b14JBIL2tnYyB7Km9y3biA+0ZvM2DU0tUUlMVKSZRavfliwY1K3T4lmeHA0NWQKlJn42t/ry1qUnf/XW2XXr4XvgBL0f9V741mWmpTZRVVs1x2Ni784bFs7MzkyXJVBm8mfTVq3p/eYWrTNTvwrUY6jLhmOXKIpKS0wIef7QqqMdALR36O6z8y/6gNiwYH1jE5YMv6bCjGRd01b0vo6pRWFmaq0/5sLMlLDAC93dvnzD0GpqlhLxujAzpaqiPPL+1fKiAj6vYaeKOamJRi2t6H0jM8u89ORav6/ctORn184MnSXTF7UGYei3YzZzJJgSHyGRHQUgAErK9v0hpPgZEnT37vD06ZfzZdrRo9CneppwY2MwNRVuzZsLaw8eBADQ1oaZMyE0FCorxRfW0rw5mJqCuTl4eICdHfz7LwAAv3oC77qfEwAwcCBs2wZjxsDq1bL0Jj8/n83h0PscLreqqqqkpER6LUEQDAbjyuVLNu2s7965s3CRtyyBigsL1Vlsel+dzeHzq8rLSkW1auosDW0dAHh+P3Cz9yz6ZDk/N+vFv3ct23XY9NcZHX3DjQunyxKoqDCfxRa+Zjabw6+qKistlf4jNIqifl+zfI73MlkOBoCiggIWm92gQLnZWf/evd2uo83h05f0DQ0XTfeQJVBxUYF6dSAWR8xbp6mtAwDPggLXL5w5/7fNAJCXnaWtpz9iorvf0bNKyipr53rKEqikqFBVXRhIjc0W8PkVNQIRBEEwGC/u3lwwvEfI03+He8wCAFV1FldLBwDePLy7a5nXr8s2yBKIV1Kkoib8YqGixiYF/MqKsi/VFHV//7oeU5eoqLNEZZZOA43b2J30Gn7k1z68kiIA4Fc2LEGXFxc1URP2TpXFEfD5leVfglIUdd5vxUivlarqnAY125goioB6R3HUTSM/zs8wDrp1a/Dygv79Yft26NwZ0tPBzw/S08Gn+oO3uBgKa9wKUFcHXV3w8oJ+/YDNhp07wc4OVFTEFwLA5cvQtSvQ51Z0OwIBPHkCL1/CunWgqQmhoRAeDiYmsH+/mJenpgYA4OsLrVvDvHnC4XeSaWlplRQX0/slxcVKSkrs6lwgvdZlzNg+ffoeO3pk6OBBL98E1/u2cTQ0ykqFqb+spITJVFKr8d8PAIoL87ev8s5KS11/wN+yXUcAUGmi2qFz1yHjpwDAzGXrhnZsUVSQz61xQioWV0OrtET4mktLS5hKSuoslvQfoYW9e8tis1u0tJDlYADgamqWVn+YlZbIFEilSRN7R6dxblMBYOlaX1szo4L8PE0tbek/xeFqlpVIe+uKCvK3rvTOSkvZ/OdJq3YdAaBdp85/33lG13pv2Dqoo1lhfp5GfYHYXI2KMmGgitJSJlNJVb12j7oNHNbB0Tno8ukNMyduu3AXAEoKCw6u98lJT122+5h52/qv3QNAEzZXlByryksZTKaK6pcLQRFBV1ja+i3snCuKC778DEH0mbW61/TlFEWVF+aHBJxRZcv0/UNEjcPllVf3rqyEwWSKPiQA4OXN8xq6Bm269iotLJDQgAIgQPo4aAGAEo6Dhq1b4bffwM8P2rYFNzcwMoI3b77cwevdGzQ1v2wXLsCIETB1KvToAVZWEBYGZ88CgPhCAPDwgPfvhfumpqCpCTo6MGcOrF8Pzs5gZQVeXuDkBL16wZw5El+hsTEsXgze9Z/btjAz//Ahmt6PiYkxNW1Rc9FCsbUnjh/7689DAKClrT1t+ozwsLC6Yz/qMjJpkfRReIkmOSHesJlJzUD8qsplU8eZtrTcfymQzs4AUPPKKcEgCAaDwaj/L8TEtEVCXAy9nxAX29SkuYzLMN4LuDHMZZwsR4oCfRQFipcpkFHTZqJ9BsFgEAwGg1lvIOPmponVb11SQlytt66qqnKx+9gWFq0OXbljVf3WRYUEh7wSXl5TVlFmMphKSvWf/Rg0M02pHj+T+iler2mzmoHuXzl758JJAGBraPYf65YYE8WvquJXVW2YObGZmcWW0zdlzM4AoGHQLC8lgd7PS/3E1W9acx2Q5LBXCW8eHnbv4T9nGCkQHHbvkREbnvM59vmpPQymElNJOS36nW4LSylLh4ilY9w8s3ogUFZigrbRV72LC34R8Sxo9VD7za59SYFg9VD7xKiQBrXfGKh6rkGTFPB5cjyD/kkSNABMnQqvXkFxMcTEwL59IFqGIDu79leOSZMAADZuhPR0KCqCwECwqD5HE1tYXAxDhgCL9VUjqalf7iLu3AlFRRASAmPHCgfSdewozOlDhsDVq8LD1q4VXhKRqkfPnnm5eQE3b1RUVOzeuWOCqytd/s/VK+npaWJrdXV1/X7/PT4+LiM9favfFttOnVRUJC7iINKxS7eigvzn9wMrebyLxw70HT6GLn9y52ZuVsaze7cFAsHgcW5Z6WmZqSmZqSkUSdp37/0pJurRrWtlpSUndm1pa2PH5tZ/0tS5W/eC/Px/79zi8SpOHNo7zEU4fvzuzetZGdIuxd67dd2pV9962xfp0s25ID/vwZ1bPF7FsYN7h48RBbomKZBTrz6x0ZG3r10pLSnZ7edrY9+ZK8NlaFsHp6L8/GdBgZU83rkjBwaMHEuXPwq8mZOZ8fTubZIUDB3vlpWempGanJGaTJIkj1exarZ7ePDr0pLiIzu2tLd3kOXSsHXnrsUF+W8e3q3i8a77/9ljiAtd/jIoIC87k6ulffnw7vSkT/nZWVf+2mPetoOSsvLrB4EkKejnMiknIy07LSU7LYWSYSHaZu06VxQXJLx+KKjivbvm37rnELo8/kVQaV5275mr3Pb847rjkovvMYLBcN1xSc+8NVvXIPTW2fiXQUWZqa/OH2o/YEy9UWqxsHUoLcqPeBrEr+Q9PHfEfsBIujz0UWBhTuYY7/XL/77rczxg7t4zDAbT53hAs1bWDQ3RCCh65W4pmzwXAv55EvR/BZPJvHD5yoZ1a1uZt+BwOD7LltPlMzx/DX0fIrZ26LDhUzw8+vfp3bFd2/DwMP9Tp2UJxGAyNxz099/tN7FHR3UW23XWArp867J58VHhcVFhH6MjJvXuJNqKiwpVmjTZcPDkmUO7JnTvkPAhasU2MeOixPZo74kze7Zu6mPblsXizFgg/BqxcsGs6AiJA05ioiIreTwDI5lGO4gC7T9xdo/fpl42bVhs9swFi+ny5fNnRUkY2dKkiep+/7N/7t7eo6NlTFTE1gN/yRKIwWRu+vPk0V1bXJw6qLPYk2cL37rNPnPjosJjI8PioiLG9egk2kqKCmwdnNznLl47z3OMU8ekj3FrdhyUKRCDuWz3sfP7/5jR316NxRo9bR5dvm/1ok/REZ17D+wzcvwaj9HzhzsnxkYv8jsAAAnR4Z8/RM4e5CDaSorqH+pLMBhDl+9+cXb/0Wn9VdRY9i7CUZV396zO+hStytZg6+izdfTZ2noEQbB19JlKyqpsjQHzfR8d8bu4yqOV08B2DU/QDAbTc/Oft4/uWjfaqYk6q9/k2XT52U0+qXFR6lxNDT0DDT0Dro4BEKChZ8CU53CIb1Q9H7SUTVlNjqt6E5TYkQnoB+HxSSkzrfxAUZnFBY2yqreOunITpcb4XGc08Av1NyusqGqcxbYLK/i8Rlk+PDqjOKdUjt+7afxK3v7xnYfOWtp3kkxjsb9Ho63qXRXq77Ns5dYtwgGmBVk5DsamU1QkDgQsoPgvtIinaZ/k9Ip+hpuECCH0v0AQIH3JK5ICgln/XY1vhgkaIYQkqHfJK4piyHA3+JthgkYIIfGE80FLPoAEqKr4wc+p14QJGiGExCMIkL7kFSnnyZIwQSOEkHgURQEFUsbRkQAUrqiCEEL/ExQAJXnCDQooSdN1ZGVCAAACgElEQVSo/hCYoBFCSAKCoIDiS7lJCJRcH/XGBI0QQhLQj3pLricBH/VGCKH/BfrMWfpcHA1a1dvX1/fy5cuyH48JGiGExJNhyStKWbZLHCRJ9urVa/16mZZDEsEEjRBCElRf4pC0UQCVZTKt1shgMO7fv7906dIGxcdr0Aihb/fgzOEX18/Wf9z3KSkuk+ul3pq+msOWIIqA/46SuKhYBZCFAlU/Pz9TU9MJEyZIb5nJZMoyf29NmKDlS5lJADTGpD+WeuyqRpmIR0CCukpjfPESkJRSA/+av40eX9A4EzOVVgqUmY0RqJ0hh2iUv7qiGXPzsmRa6e07lZXxBAJ+I3SJIIhxY0aL/qmpq7N3957KColLyVAUxWOCpU17XV1dsQecOHEiICDAxsZm1apV3/J6cDY7hBBqHGvWrLGxsXFxcZHxeLwGjRBCCgrPoBFCSEHhGTRCCCkoTNAIIaSgMEEjhJCCwgSNEEIKChM0QggpKEzQCCGkoDBBI4SQgsIEjRBCCgoTNEIIKShM0AghpKAwQSOEkILCBI0QQgoKEzRCCCkoTNAIIaSgMEEjhJCCwgSNEEIKChM0QggpKEzQCCGkoDBBI4SQgsIEjRBCCgoTNEIIKShM0AghpKAwQSOEkILCBI0QQgoKEzRCCCkoTNAIIaSgMEEjhJCCwgSNEEIKChM0QggpKEzQCCGkoDBBI4SQgsIEjRBCCgoTNEIIKShM0AghpKAwQSOEkILCBI0QQgoKEzRCCCkoTNAIIaSgMEEjhJCCwgSNEEIKChM0QggpKEzQCCGkoDBBI4SQgsIEjRBCCgoTNEIIKShM0AghpKAwQSOEkILCBI0QQgrq/wCnhtX+NNv8+AAAAABJRU5ErkJggg==\n"
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%%R\n",
    "corrplot(M, method=\"color\", p.mat = M, insig = \"p-value\", sig.level=-1, order=\"hclust\", addrect=3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 283,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "png \n",
       "  2 \n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%%R\n",
    "pdf('corHeat_updated.pdf')\n",
    "corrplot(M, method=\"color\", p.mat = M, insig = \"p-value\", sig.level=-1, order=\"hclust\", addrect=3)\n",
    "dev.off()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The correlation coefficients are printed in each cell. we set the `sig.level` to -1 so that all the values are printed in the plot. From the plot it we could assign Cluster1 with OE33, Cluster2 with  OE19 and Cluster3 with HET1A. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Differential Accessibility (Cluster1 vs Cluster2)\n",
    "We now calculate Differential Accessibility to find out the peaks that are differentially accessible in Cluster1 and Cluster2. As we know that cells in these two groups are associated with OE33 and OE19 respectively, the expressed genes are also very similar. So it is interesting to find out the genes that are differentially expressed in these two cells."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We first extract the cells in each cluster and their count data."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 134,
   "metadata": {},
   "outputs": [],
   "source": [
    "%%R\n",
    "c1CellName <- names(pamx$clustering[pamx$clustering == 1])\n",
    "c2CellName <-  names(pamx$clustering[pamx$clustering == 2])\n",
    "c3CellName <-  names(pamx$clustering[pamx$clustering == 3])\n",
    "c1CountData <- cdFiltTopSel[,as.character(c1CellName)]\n",
    "colnames(c1CountData) <- paste0('C1_', colnames(c1CountData))\n",
    "c2CountData <- cdFiltTopSel[,as.character(c2CellName)]\n",
    "colnames(c2CountData) <- paste0('C2_', colnames(c2CountData))\n",
    "c3CountData <- cdFiltTopSel[,as.character(c3CellName)]\n",
    "colnames(c3CountData) <- paste0('C3_', colnames(c3CountData))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We prepare the data for DE analysis"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 135,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "sg\n",
       "C1 C2 \n",
       "41 80 \n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%%R\n",
    "CellCluster.Annot.counts <- cbind(cdFiltTopSel[,1:3], c1CountData ,c2CountData)\n",
    "rownames(CellCluster.Annot.counts) <- paste0(cdFiltTopSel[,1],'_',cdFiltTopSel[,2],'_',cdFiltTopSel[,3])\n",
    "sg <- factor(gsub(\"(C1|C2).*\", \"\\\\1\", colnames(CellCluster.Annot.counts[,4:dim(CellCluster.Annot.counts)[2]])), \n",
    "             levels = c(\"C1\", \"C2\"))\n",
    "# the group factor should be named accordingly\n",
    "names(sg) <- colnames(CellCluster.Annot.counts[,4:dim(CellCluster.Annot.counts)[2]]) \n",
    "table(sg)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 136,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Time difference of 2.1992 mins\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%%R\n",
    "start.time <- Sys.time()\n",
    "diffRes <- getDiffAccessInformationGain(CellCluster.Annot.counts, groups=sg)\n",
    "end.time <- Sys.time()\n",
    "time.taken <- end.time - start.time\n",
    "time.taken"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We consider all the peaks that have an information gain of above 0.01 to be significant. So we extract those peaks."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 137,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[1] 12583     7\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%%R\n",
    "significantDEpeaks <- diffRes\n",
    "significantDEpeaks <- significantDEpeaks[order(significantDEpeaks$information.gain, decreasing=TRUE),]\n",
    "significantDEpeaks <- significantDEpeaks[significantDEpeaks$information.gain>0.01,]\n",
    "rownames(significantDEpeaks) <- paste0(significantDEpeaks[,1],'_',significantDEpeaks[,2],'_',significantDEpeaks[,3])\n",
    "print(dim(significantDEpeaks))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "There are 5365 peaks that are significantly differentially accessible in OE19 and OE33."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 138,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "                            Chr     Start       end information.gain\n",
       "chr17_39535097_39537584   chr17  39535097  39537584        0.2671008\n",
       "chr9_37963667_37965216     chr9  37963667  37965216        0.2318043\n",
       "chr5_912358_913095         chr5    912358    913095        0.2202554\n",
       "chr17_35846992_35851866   chr17  35846992  35851866        0.2178187\n",
       "chr10_106908425_106909662 chr10 106908425 106909662        0.2086013\n",
       "chr13_30646183_30647194   chr13  30646183  30647194        0.1977572\n",
       "                          rawMeanGroup1 rawMeanGroup2 log2FoldChange\n",
       "chr17_39535097_39537584      0.70731707        0.0375       4.237395\n",
       "chr9_37963667_37965216       0.07317073        0.7500      -3.357552\n",
       "chr5_912358_913095           0.65853659        0.0500       3.719264\n",
       "chr17_35846992_35851866      0.87804878        0.2125       2.046838\n",
       "chr10_106908425_106909662    0.60975610        0.0375       4.023270\n",
       "chr13_30646183_30647194      0.02439024        0.6125      -4.650334\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%%R\n",
    "head(significantDEpeaks)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We identify the peaks that are highly open in OE33 cells."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 143,
   "metadata": {},
   "outputs": [],
   "source": [
    "%%R\n",
    "sigOE33 <- significantDEpeaks[significantDEpeaks$log2FoldChange>0,]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Annotating peaks\n",
    "\n",
    "We annotate the genes associated with the peaks now"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Converting the peaks to GRanges object"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 144,
   "metadata": {},
   "outputs": [],
   "source": [
    "%%R\n",
    "gr1 <- makeGRangesFromDataFrame(as.data.frame(sigOE33[,1:3]),\n",
    "                         keep.extra.columns=FALSE,\n",
    "                         ignore.strand=TRUE,\n",
    "                         seqinfo=NULL,\n",
    "                         seqnames.field=\"Chr\",\n",
    "                         #start.field=\"start\",\n",
    "                         #end.field=c(\"end\", \"stop\"),                         \n",
    "                         starts.in.df.are.0based=FALSE)\n",
    "metaData <- data.frame(peakNames = rownames(sigOE33), stringsAsFactors=FALSE)\n",
    "mcols(gr1) <- metaData"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Prepare annotation data\n",
    "\n",
    "Annotation data should be an object of GRanges. We use the method toGRanges, which can return an object of GRanges, to represent the annotation data. An annotation data be constructed from not only BED, GFF or user defined readable text files, but also EnsDb or TxDb object, by calling the toGRanges method. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 145,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "GRanges object with 2 ranges and 1 metadata column:\n",
       "                  seqnames         ranges strand |   gene_name\n",
       "                     <Rle>      <IRanges>  <Rle> | <character>\n",
       "  ENSG00000223972     chr1 [11869, 14412]      + |     DDX11L1\n",
       "  ENSG00000227232     chr1 [14363, 29806]      - |      WASH7P\n",
       "  -------\n",
       "  seqinfo: 273 sequences from GRCh37 genome\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%%R\n",
    "annoData <- toGRanges(EnsDb.Hsapiens.v75, feature=\"gene\")\n",
    "annoData[1:2,'gene_name']"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Annotate peaks\n",
    "\n",
    "Now we will use `annotatePeakInBatch()` or `annoPeaks()` to annotate the peaks to the overlaping regions of Hg19 genes."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 146,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[1] \"KRT34\"  \"TRIP13\" NA       \"SORCS3\" \"DIP2C\"  NA      \n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%%R\n",
    "gr1.anno <- annotatePeakInBatch(gr1, \n",
    "                                     AnnotationData=annoData, \n",
    "                                     #output=\"nearestBiDirectionalPromoters\",\n",
    "                                     output = \"overlapping\",\n",
    "                                     #bindingRegion=c(-2000, 500))\n",
    "                                     maxgap=100)\n",
    "gr1.anno <- addGeneIDs(gr1.anno,\n",
    "                            \"org.Hs.eg.db\",\n",
    "                            IDs2Add = \"symbol\")\n",
    "print(head(gr1.anno$symbol))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 147,
   "metadata": {},
   "outputs": [],
   "source": [
    "%%R\n",
    "write.csv(gr1.anno$symbol, 'OE33_Significant.csv')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "One gene that have differential accessibillity in OE33 cells are _CAV1_ for which we visualize the peaks associated with this gene location in UCSCS browser which is reported in the paper. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Differential Accessibility (Cluster2 vs Cluster3)\n",
    "We now calculate Differential Accessibility to find out the peaks that are differentially accessible in Cluster2 and Cluster3. As we know that cells in these two groups are associated with OE19 and HET1A respectively, the expressed genes are also very similar. So it is interesting to find out the genes that are differentially expressed in these two cells."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We first extract the cells in each cluster and their count data."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We prepare the data for DE analysis"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 148,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "sg\n",
       "C2 C3 \n",
       "80 52 \n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%%R\n",
    "CellCluster.Annot.counts <- cbind(cdFiltTopSel[,1:3], c2CountData ,c3CountData)\n",
    "rownames(CellCluster.Annot.counts) <- paste0(cdFiltTopSel[,1],'_',cdFiltTopSel[,2],'_',cdFiltTopSel[,3])\n",
    "sg <- factor(gsub(\"(C2|C3).*\", \"\\\\1\", colnames(CellCluster.Annot.counts[,4:dim(CellCluster.Annot.counts)[2]])), \n",
    "             levels = c(\"C2\", \"C3\"))\n",
    "# the group factor should be named accordingly\n",
    "names(sg) <- colnames(CellCluster.Annot.counts[,4:dim(CellCluster.Annot.counts)[2]]) \n",
    "table(sg)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 149,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Time difference of 2.174719 mins\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%%R\n",
    "start.time <- Sys.time()\n",
    "diffRes <- getDiffAccessInformationGain(CellCluster.Annot.counts, groups=sg)\n",
    "end.time <- Sys.time()\n",
    "time.taken <- end.time - start.time\n",
    "time.taken"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We consider all the peaks that have an information gain of above 0.01 to be significant. So we extract those peaks."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 192,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[1] 15391     7\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%%R\n",
    "significantDEpeaks <- diffRes\n",
    "significantDEpeaks <- significantDEpeaks[order(significantDEpeaks$information.gain, decreasing=TRUE),]\n",
    "significantDEpeaks <- significantDEpeaks[significantDEpeaks$information.gain>0.01,]\n",
    "rownames(significantDEpeaks) <- paste0(significantDEpeaks[,1],'_',significantDEpeaks[,2],'_',significantDEpeaks[,3])\n",
    "print(dim(significantDEpeaks))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "There are 15391 peaks that are significantly differentially accessible in OE19 and HET1A."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 193,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "                           Chr     Start       end information.gain\n",
       "chr17_37858661_37867115  chr17  37858661  37867115        0.3570128\n",
       "chr17_37852761_37858463  chr17  37852761  37858463        0.3418310\n",
       "chr17_39944768_39950869  chr17  39944768  39950869        0.3070466\n",
       "chr7_157101284_157103548  chr7 157101284 157103548        0.2862330\n",
       "chr9_71354539_71356329    chr9  71354539  71356329        0.2683219\n",
       "chr6_42084948_42087120    chr6  42084948  42087120        0.2631854\n",
       "                         rawMeanGroup1 rawMeanGroup2 log2FoldChange\n",
       "chr17_37858661_37867115         0.9750    0.21153846       2.204482\n",
       "chr17_37852761_37858463         0.9750    0.23076923       2.078951\n",
       "chr17_39944768_39950869         0.9625    0.25000000       1.944858\n",
       "chr7_157101284_157103548        0.0250    0.69230769      -4.791413\n",
       "chr9_71354539_71356329          0.7250    0.03846154       4.236493\n",
       "chr6_42084948_42087120          0.7750    0.07692308       3.332708\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%%R\n",
    "head(significantDEpeaks)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We identify the peaks that are highly open in OE19 cells."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 206,
   "metadata": {},
   "outputs": [],
   "source": [
    "%%R\n",
    "sigOE19 <- significantDEpeaks[significantDEpeaks$log2FoldChange>0,]\n",
    "sigHET1A <- significantDEpeaks[significantDEpeaks$log2FoldChange<0,]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Annotating peaks\n",
    "\n",
    "We annotate the genes associated with the peaks now"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Converting the peaks to GRanges object"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 207,
   "metadata": {},
   "outputs": [],
   "source": [
    "%%R\n",
    "gr1 <- makeGRangesFromDataFrame(as.data.frame(sigOE19[,1:3]),\n",
    "                         keep.extra.columns=FALSE,\n",
    "                         ignore.strand=TRUE,\n",
    "                         seqinfo=NULL,\n",
    "                         seqnames.field=\"Chr\",\n",
    "                         #start.field=\"start\",\n",
    "                         #end.field=c(\"end\", \"stop\"),                         \n",
    "                         starts.in.df.are.0based=FALSE)\n",
    "metaData <- data.frame(peakNames = rownames(sigOE19), stringsAsFactors=FALSE)\n",
    "mcols(gr1) <- metaData"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Prepare annotation data\n",
    "\n",
    "Annotation data should be an object of GRanges. We use the method toGRanges, which can return an object of GRanges, to represent the annotation data. An annotation data be constructed from not only BED, GFF or user defined readable text files, but also EnsDb or TxDb object, by calling the toGRanges method. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 208,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "GRanges object with 2 ranges and 1 metadata column:\n",
       "                  seqnames         ranges strand |   gene_name\n",
       "                     <Rle>      <IRanges>  <Rle> | <character>\n",
       "  ENSG00000223972     chr1 [11869, 14409]      + |     DDX11L1\n",
       "  ENSG00000227232     chr1 [14404, 29570]      - |      WASH7P\n",
       "  -------\n",
       "  seqinfo: 357 sequences from GRCh38 genome\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%%R\n",
    "annoData <- toGRanges(EnsDb.Hsapiens.v86, feature=\"gene\")\n",
    "annoData[1:2,'gene_name']"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Annotate peaks\n",
    "\n",
    "Now we will use `annotatePeakInBatch()` or `annoPeaks()` to annotate the peaks to the overlaping regions of Hg19 genes."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 209,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[1] \"YWHAEP7\" \"YWHAEP7\" \"LRRC3C\"  \"TRPM3\"   \"TAF8\"    NA       \n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%%R\n",
    "gr1.anno <- annotatePeakInBatch(gr1, \n",
    "                                     AnnotationData=annoData, \n",
    "                                     #output=\"nearestBiDirectionalPromoters\",\n",
    "                                     output = \"overlapping\",\n",
    "                                     #bindingRegion=c(-2000, 500))\n",
    "                                     maxgap=100)\n",
    "gr1.anno <- addGeneIDs(gr1.anno,\n",
    "                            \"org.Hs.eg.db\",\n",
    "                            IDs2Add = \"symbol\")\n",
    "print(head(gr1.anno$symbol))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 210,
   "metadata": {},
   "outputs": [],
   "source": [
    "%%R\n",
    "write.csv(gr1.anno$symbol, 'OE19_Significant.csv')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Two genes reported in the literature that are differentially expressed in OE19 are _GATA6_ and _PPFIA3_. Peaks associated with these two genes also have differential accessibility in our OE19 significant peaks. We then visualize the peaks of these two gene location in UCSCS browser which is reported in the paper. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Disease ontology analysis\n",
    "\n",
    "We now do a disease ontology analysis with the OE19 significant peaks and HET1A significant peaks. We will first identify the disease ontoogy assocaited with OE19 cells and then will try to find the significance of same disease ontology on HET1A cells. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 211,
   "metadata": {},
   "outputs": [],
   "source": [
    "%%R\n",
    "singOE19Sel <- sigOE19[sigOE19$log2FoldChange>2,]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 212,
   "metadata": {},
   "outputs": [],
   "source": [
    "%%R\n",
    "bedg1 = singOE19Sel\n",
    "jobg1 = submitGreatJob(bedg1, version = \"3.0\", species='hg19')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 213,
   "metadata": {},
   "outputs": [],
   "source": [
    "%%R -w 550 -h 350\n",
    "tbDisG1 = getEnrichmentTables(jobg1, ontolog=\"Disease Ontology\")\n",
    "dftbDisG1 = data.frame(Name = tbDisG1[[1]][,2], Binom_p=-log10(tbDisG1[[1]][,8]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 214,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "                                   Name  Binom_p\n",
       "1                                cancer 35.66909\n",
       "2                             carcinoma 31.02175\n",
       "3                  epithelial carcinoma 30.08280\n",
       "4            reproductive system cancer 25.23526\n",
       "5          malignant neoplasm of thorax 24.58818\n",
       "6          malignant neoplasm of breast 24.49216\n",
       "7                       thoracic cancer 24.07995\n",
       "8           cancer by anatomical entity 23.91215\n",
       "9                        mammary cancer 23.10336\n",
       "10                       adenocarcinoma 22.89754\n",
       "11                              disease 22.38073\n",
       "12                        neck neoplasm 21.70345\n",
       "13                          neck cancer 21.61245\n",
       "14         disease of anatomical entity 20.83940\n",
       "15   glandular cell epithelial neoplasm 20.60640\n",
       "16               squamous cell neoplasm 20.44316\n",
       "17    glandular and epithelial neoplasm 19.48848\n",
       "18                      T-cell leukemia 19.46805\n",
       "19                       POEMS syndrome 19.37241\n",
       "20                 neovascular glaucoma 19.37241\n",
       "21 mucinous adenocarcinoma of the ovary 19.37241\n",
       "22                cystoid macular edema 19.37241\n",
       "23                        macular holes 19.37241\n",
       "24        cholesteatoma of external ear 19.37241\n",
       "25          transitional cell carcinoma 18.97086\n",
       "26                 genitourinary cancer 18.87226\n",
       "27                     thyroid neoplasm 18.86292\n",
       "28          active peptic ulcer disease 18.66238\n",
       "29        malignant neoplasm of thyroid 18.18632\n",
       "30                    laryngeal disease 17.74633\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%%R\n",
    "dftbDisG1[1:30,]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 217,
   "metadata": {},
   "outputs": [],
   "source": [
    "%%R\n",
    "sigHET1ASel <- sigHET1A[sigHET1A$log2FoldChange < (-2),]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 218,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/mqbsxsm2/anaconda3/lib/python3.6/site-packages/rpy2/rinterface/__init__.py:145: RRuntimeWarning: Don't make too frequent requests. The time break is 300s.\n",
      "Please wait for 143s for the next request.\n",
      "The time break can be set by `request_interval` argument.\n",
      "\n",
      "\n",
      "  warnings.warn(x, RRuntimeWarning)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "\r",
       "  |                                                                            \r",
       "  |                                                                      |   0%\r",
       "  |                                                                            \r",
       "  |                                                                      |   1%\r",
       "  |                                                                            \r",
       "  |=                                                                     |   1%\r",
       "  |                                                                            \r",
       "  |=                                                                     |   2%\r",
       "  |                                                                            \r",
       "  |==                                                                    |   3%\r",
       "  |                                                                            \r",
       "  |===                                                                   |   4%\r",
       "  |                                                                            \r",
       "  |===                                                                   |   5%\r",
       "  |                                                                            \r",
       "  |====                                                                  |   6%\r",
       "  |                                                                            \r",
       "  |=====                                                                 |   7%\r",
       "  |                                                                            \r",
       "  |=====                                                                 |   8%\r",
       "  |                                                                            \r",
       "  |======                                                                |   8%\r",
       "  |                                                                            \r",
       "  |======                                                                |   9%\r",
       "  |                                                                            \r",
       "  |=======                                                               |  10%\r",
       "  |                                                                            \r",
       "  |========                                                              |  11%\r",
       "  |                                                                            \r",
       "  |========                                                              |  12%\r",
       "  |                                                                            \r",
       "  |=========                                                             |  13%\r",
       "  |                                                                            \r",
       "  |==========                                                            |  14%\r",
       "  |                                                                            \r",
       "  |==========                                                            |  15%\r",
       "  |                                                                            \r",
       "  |===========                                                           |  15%\r",
       "  |                                                                            \r",
       "  |===========                                                           |  16%\r",
       "  |                                                                            \r",
       "  |============                                                          |  17%\r",
       "  |                                                                            \r",
       "  |=============                                                         |  18%\r",
       "  |                                                                            \r",
       "  |=============                                                         |  19%\r",
       "  |                                                                            \r",
       "  |==============                                                        |  20%\r",
       "  |                                                                            \r",
       "  |===============                                                       |  21%\r",
       "  |                                                                            \r",
       "  |===============                                                       |  22%\r",
       "  |                                                                            \r",
       "  |================                                                      |  22%\r",
       "  |                                                                            \r",
       "  |================                                                      |  23%\r",
       "  |                                                                            \r",
       "  |=================                                                     |  24%\r",
       "  |                                                                            \r",
       "  |==================                                                    |  25%\r",
       "  |                                                                            \r",
       "  |==================                                                    |  26%\r",
       "  |                                                                            \r",
       "  |===================                                                   |  27%\r",
       "  |                                                                            \r",
       "  |====================                                                  |  28%\r",
       "  |                                                                            \r",
       "  |====================                                                  |  29%\r",
       "  |                                                                            \r",
       "  |=====================                                                 |  29%\r",
       "  |                                                                            \r",
       "  |=====================                                                 |  30%\r",
       "  |                                                                            \r",
       "  |======================                                                |  31%\r",
       "  |                                                                            \r",
       "  |=======================                                               |  32%\r",
       "  |                                                                            \r",
       "  |=======================                                               |  33%\r",
       "  |                                                                            \r",
       "  |=======================                                               |  34%\r",
       "  |                                                                            \r",
       "  |========================                                              |  34%\r",
       "  |                                                                            \r",
       "  |========================                                              |  35%\r",
       "  |                                                                            \r",
       "  |=========================                                             |  36%\r",
       "  |                                                                            \r",
       "  |==========================                                            |  37%\r",
       "  |                                                                            \r",
       "  |==========================                                            |  38%\r",
       "  |                                                                            \r",
       "  |===========================                                           |  38%\r",
       "  |                                                                            \r",
       "  |===========================                                           |  39%\r",
       "  |                                                                            \r",
       "  |============================                                          |  40%\r",
       "  |                                                                            \r",
       "  |============================                                          |  41%\r",
       "  |                                                                            \r",
       "  |=============================                                         |  41%\r",
       "  |                                                                            \r",
       "  |=============================                                         |  42%\r",
       "  |                                                                            \r",
       "  |==============================                                        |  43%\r",
       "  |                                                                            \r",
       "  |===============================                                       |  44%\r",
       "  |                                                                            \r",
       "  |===============================                                       |  45%\r",
       "  |                                                                            \r",
       "  |================================                                      |  45%\r",
       "  |                                                                            \r",
       "  |================================                                      |  46%\r",
       "  |                                                                            \r",
       "  |=================================                                     |  47%\r",
       "  |                                                                            \r",
       "  |=================================                                     |  48%\r",
       "  |                                                                            \r",
       "  |==================================                                    |  48%\r",
       "  |                                                                            \r",
       "  |==================================                                    |  49%\r",
       "  |                                                                            \r",
       "  |===================================                                   |  50%\r",
       "  |                                                                            \r",
       "  |====================================                                  |  51%\r",
       "  |                                                                            \r",
       "  |====================================                                  |  52%\r",
       "  |                                                                            \r",
       "  |=====================================                                 |  52%\r",
       "  |                                                                            \r",
       "  |=====================================                                 |  53%\r",
       "  |                                                                            \r",
       "  |======================================                                |  54%\r",
       "  |                                                                            \r",
       "  |======================================                                |  55%\r",
       "  |                                                                            \r",
       "  |=======================================                               |  55%\r",
       "  |                                                                            \r",
       "  |=======================================                               |  56%\r",
       "  |                                                                            \r",
       "  |========================================                              |  57%\r",
       "  |                                                                            \r",
       "  |=========================================                             |  58%\r",
       "  |                                                                            \r",
       "  |=========================================                             |  59%\r",
       "  |                                                                            \r",
       "  |==========================================                            |  59%\r",
       "  |                                                                            \r",
       "  |==========================================                            |  60%\r",
       "  |                                                                            \r",
       "  |===========================================                           |  61%\r",
       "  |                                                                            \r",
       "  |===========================================                           |  62%\r",
       "  |                                                                            \r",
       "  |============================================                          |  62%\r",
       "  |                                                                            \r",
       "  |============================================                          |  63%\r",
       "  |                                                                            \r",
       "  |=============================================                         |  64%\r",
       "  |                                                                            \r",
       "  |==============================================                        |  65%\r",
       "  |                                                                            \r",
       "  |==============================================                        |  66%\r",
       "  |                                                                            \r",
       "  |===============================================                       |  67%\r",
       "  |                                                                            \r",
       "  |===============================================                       |  68%\r",
       "  |                                                                            \r",
       "  |================================================                      |  69%\r",
       "  |                                                                            \r",
       "  |=================================================                     |  70%\r",
       "  |                                                                            \r",
       "  |=================================================                     |  71%\r",
       "  |                                                                            \r",
       "  |==================================================                    |  71%\r",
       "  |                                                                            \r",
       "  |==================================================                    |  72%\r",
       "  |                                                                            \r",
       "  |===================================================                   |  73%\r",
       "  |                                                                            \r",
       "  |====================================================                  |  74%\r",
       "  |                                                                            \r",
       "  |====================================================                  |  75%\r",
       "  |                                                                            \r",
       "  |=====================================================                 |  76%\r",
       "  |                                                                            \r",
       "  |======================================================                |  77%\r",
       "  |                                                                            \r",
       "  |======================================================                |  78%\r",
       "  |                                                                            \r",
       "  |=======================================================               |  78%\r",
       "  |                                                                            \r",
       "  |=======================================================               |  79%\r",
       "  |                                                                            \r",
       "  |========================================================              |  80%\r",
       "  |                                                                            \r",
       "  |=========================================================             |  81%\r",
       "  |                                                                            \r",
       "  |=========================================================             |  82%\r",
       "  |                                                                            \r",
       "  |==========================================================            |  83%\r",
       "  |                                                                            \r",
       "  |===========================================================           |  84%\r",
       "  |                                                                            \r",
       "  |===========================================================           |  85%\r",
       "  |                                                                            \r",
       "  |============================================================          |  85%\r",
       "  |                                                                            \r",
       "  |============================================================          |  86%\r",
       "  |                                                                            \r",
       "  |=============================================================         |  87%\r",
       "  |                                                                            \r",
       "  |==============================================================        |  88%\r",
       "  |                                                                            \r",
       "  |==============================================================        |  89%\r",
       "  |                                                                            \r",
       "  |===============================================================       |  89%\r",
       "  |                                                                            \r",
       "  |===============================================================       |  90%\r",
       "  |                                                                            \r",
       "  |================================================================      |  91%\r",
       "  |                                                                            \r",
       "  |================================================================      |  92%\r",
       "  |                                                                            \r",
       "  |=================================================================     |  92%\r",
       "  |                                                                            \r",
       "  |=================================================================     |  93%\r",
       "  |                                                                            \r",
       "  |==================================================================    |  94%\r",
       "  |                                                                            \r",
       "  |===================================================================   |  95%\r",
       "  |                                                                            \r",
       "  |===================================================================   |  96%\r",
       "  |                                                                            \r",
       "  |====================================================================  |  96%\r",
       "  |                                                                            \r",
       "  |====================================================================  |  97%\r",
       "  |                                                                            \r",
       "  |===================================================================== |  98%\r",
       "  |                                                                            \r",
       "  |===================================================================== |  99%\r",
       "  |                                                                            \r",
       "  |======================================================================| 100%\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%%R\n",
    "bedg2 = sigHET1A[,1:3]\n",
    "jobg2 = submitGreatJob(bedg2, version = \"3.0\", species='hg19')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 219,
   "metadata": {},
   "outputs": [],
   "source": [
    "%%R -w 550 -h 350\n",
    "tbDisG2 = getEnrichmentTables(jobg2, ontolog=\"Disease Ontology\")\n",
    "dftbDisG2 = data.frame(Name = tbDisG2[[1]][,2], Binom_p=-log10(tbDisG2[[1]][,8]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 220,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "                                         Name  Binom_p\n",
       "1              dsDNA virus infectious disease 22.89563\n",
       "2                DNA virus infectious disease 20.41672\n",
       "3                  reproductive system cancer 19.68900\n",
       "4                              mammary cancer 19.52828\n",
       "5  progressive multifocal leukoencephalopathy 18.72189\n",
       "6            Herpesviridae infectious disease 18.06720\n",
       "7                       female genital cancer 17.47434\n",
       "8                malignant neoplasm of breast 17.24607\n",
       "9          Cytomegalovirus infectious disease 17.24238\n",
       "10               malignant neoplasm of thorax 17.12734\n",
       "11                            thoracic cancer 16.54006\n",
       "12                                     cancer 16.09954\n",
       "13                           ovarian neoplasm 16.07262\n",
       "14                     squamous cell neoplasm 15.36043\n",
       "15                   viral infectious disease 15.35467\n",
       "16                 female reproductive cancer 15.25368\n",
       "17                 neoplasm of body of uterus 15.08738\n",
       "18                disease by infectious agent 14.62749\n",
       "19                                epithelioma 14.45157\n",
       "20                    squamous cell carcinoma 14.27593\n",
       "21      head and neck squamous cell carcinoma 14.11822\n",
       "22                malignant neoplasm of ovary 13.76721\n",
       "23                             lung carcinoma 13.65348\n",
       "24 malignant neoplasm of female genital organ 13.52401\n",
       "25                               angiosarcoma 13.43817\n",
       "26                            uterine disease 13.15064\n",
       "27           pancreatic ductal adenocarcinoma 13.02966\n",
       "28               hematopoietic system disease 12.99064\n",
       "29                    pancreas adenocarcinoma 12.88269\n",
       "30                  bone inflammation disease 12.87863\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%%R\n",
    "dftbDisG2[1:30,]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 221,
   "metadata": {},
   "outputs": [],
   "source": [
    "%%R\n",
    "topG1 <- dftbDisG1[1:15,]\n",
    "#Adding the esophageal carcinoma as well as OE19 cells are esophageal carcinoma\n",
    "topG1 <- rbind(topG1,dftbDisG1[grep('esophageal carcinoma',dftbDisG1$Name),])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 222,
   "metadata": {},
   "outputs": [],
   "source": [
    "%%R\n",
    "clstTopInter <- dftbDisG2[as.character(dftbDisG2$Name) %in% as.character(topG1$Name),]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 223,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "                                  Name   Binom_p\n",
       "3           reproductive system cancer 19.689002\n",
       "4                       mammary cancer 19.528279\n",
       "8         malignant neoplasm of breast 17.246065\n",
       "10        malignant neoplasm of thorax 17.127343\n",
       "11                     thoracic cancer 16.540064\n",
       "12                              cancer 16.099542\n",
       "33                      adenocarcinoma 12.214813\n",
       "41         cancer by anatomical entity 11.160126\n",
       "45                epithelial carcinoma 10.755831\n",
       "47  glandular cell epithelial neoplasm 10.730023\n",
       "136       disease of anatomical entity  6.756053\n",
       "141                          carcinoma  6.627016\n",
       "216                        neck cancer  4.936358\n",
       "223                      neck neoplasm  4.889893\n",
       "294                            disease  4.056264\n",
       "313               esophageal carcinoma  3.866564\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%%R\n",
    "clstTopInter"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 224,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "                                 Name  Cluster2  Cluster3\n",
       "1                              cancer 35.669090 19.689002\n",
       "2                           carcinoma 31.021753 19.528279\n",
       "3                epithelial carcinoma 30.082802 17.246065\n",
       "4          reproductive system cancer 25.235257 17.127343\n",
       "5        malignant neoplasm of thorax 24.588176 16.540064\n",
       "6        malignant neoplasm of breast 24.492162 16.099542\n",
       "7                     thoracic cancer 24.079950 12.214813\n",
       "8         cancer by anatomical entity 23.912145 11.160126\n",
       "9                      mammary cancer 23.103356 10.755831\n",
       "10                     adenocarcinoma 22.897540 10.730023\n",
       "11                            disease 22.380727  6.756053\n",
       "12                      neck neoplasm 21.703447  6.627016\n",
       "13                        neck cancer 21.612454  0.981200\n",
       "14       disease of anatomical entity 20.839405  4.889893\n",
       "15 glandular cell epithelial neoplasm 20.606401  4.056264\n",
       "16               esophageal carcinoma  6.914241  3.866564\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%%R\n",
    "dbBoth <- data.frame(Name=topG1$Name, Cluster2=topG1$Binom_p, Cluster3=clstTopInter$Binom_p)\n",
    "dbBoth[13,3] = 0.9812\n",
    "dbBoth"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 225,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n"
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%%R -w 600 -h 350\n",
    "library(reshape)\n",
    "df.long<-melt(dbBoth)\n",
    "p <- ggplot(df.long,aes(Name,value,fill=variable)) + ggtitle('Cluster2 & Cluster3 disease ontology') +\n",
    "    geom_bar(stat=\"identity\",position=\"dodge\",width = 1) + coord_flip()+\n",
    "        theme_light(base_size=12) +\n",
    "        theme(text=element_text(size=14,  face=\"bold\"),\n",
    "            strip.background = element_blank(),\n",
    "              panel.border     = element_blank(),\n",
    "              plot.title = element_text(hjust = 0.5))+ xlab(\"Disease Ontology\") + \n",
    "             ylab(\"-log10(Binom_p-value)\") +  \n",
    "             scale_fill_manual(\"variable\", values = c(\"Cluster2\" = \"red\", \"Cluster3\" = \"blue\"))\n",
    "print(p)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 305,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "png \n",
       "  2 \n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%%R -w 600 -h 350\n",
    "pdf('Disease_ontology_updated.pdf', width=10, height=5)\n",
    "#library(reshape)\n",
    "df.long<-melt(dbBoth)\n",
    "p <- ggplot(df.long,aes(Name,value,fill=variable)) + ggtitle('Cluster2 & Cluster3 disease ontology') +\n",
    "    geom_bar(stat=\"identity\",position=\"dodge\",width = 1) + coord_flip()+\n",
    "        theme_light(base_size=12) +\n",
    "        theme(text=element_text(size=14,  face=\"bold\"),\n",
    "            strip.background = element_blank(),\n",
    "              panel.border     = element_blank(),\n",
    "              plot.title = element_text(hjust = 0.5))+ xlab(\"Disease Ontology\") + \n",
    "             ylab(\"-log10(Binom_p-value)\") +  \n",
    "             scale_fill_manual(\"variable\", values = c(\"Cluster2\" = \"red\", \"Cluster3\" = \"blue\"))\n",
    "print(p)\n",
    "dev.off()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Recall vs precision\n",
    "\n",
    "Scasat also has a functionality to plot the recall vs precision. Here we are doing it for class2 cells agains Bulk OE19 to see the sensitivity or TRUE positive rate for peaks that are differentially accessible in OE19 vs HET1A with statistical significance."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 229,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n"
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%%R\n",
    "OE19_peaks <- read.table('/home/mqbsxsm2/Rna-seq_Data-Analysis/syed_scATAC_ReDoing/AllCluster_Bulk_Files/OE19.Bulk_peaks.narrowPeak',\n",
    "                        sep='\\t', header=FALSE)\n",
    "\n",
    "\n",
    "colnames(OE19_peaks) <- c('chr','start','end')\n",
    "#pdf('Recall_vs_precision_OE19_Clust2.pdf')\n",
    "getPeakVsPrecision(OE19_peaks[,1:3], sigOE19[,1:3], stepSize=50)\n",
    "#dev.off()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 270,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "png \n",
       "  2 \n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%%R\n",
    "pdf('Recall_vs_precision_OE19_Clust2.pdf')\n",
    "getPeakVsPrecision(OE19_peaks[,1:3], sigOE19[,1:3], stepSize=50)\n",
    "dev.off()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 230,
   "metadata": {},
   "outputs": [],
   "source": [
    "%%R\n",
    "save.image('ScAsAT_Deconv_Redo.RData')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "anaconda-cloud": {},
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}