Add runSCENIC function to sclet

.Rbuildignore

@@ -1,2 +1,3 @@
 Makefile
 bookdown
+^data-raw$

DESCRIPTION

@@ -32,8 +32,13 @@ Imports:
 Suggests:
     CellChat,
     ggsc,
+    RcisTarget,
+    SCENIC,
     Seurat
 License: Artistic-2.0
 Encoding: UTF-8
 Roxygen: list(markdown = TRUE)
 RoxygenNote: 7.3.2
+Depends: 
+    R (>= 2.10)
+LazyData: true

NAMESPACE

@@ -30,6 +30,7 @@ export(plot_list)
 export(rowData)
 export(runCellChat)
 export(runPCA)
+export(runSCENIC)
 export(subset_cell)
 export(subset_feature)
 importFrom(DropletUtils,read10xCounts)
@@ -86,6 +87,7 @@ importFrom(tinytable,group_tt)
 importFrom(tinytable,save_tt)
 importFrom(tinytable,style_tt)
 importFrom(tinytable,tt)
+importFrom(utils,data)
 importFrom(yulab.utils,get_fun_from_pkg)
 importFrom(yulab.utils,install_zip_gh)
 importFrom(yulab.utils,is.installed)

R/data_exports.R

@@ -0,0 +1,47 @@
+#' Documentation for the data
+
+#' Default databases by organism
+#' 
+#' @format Named character vector
+#' @source GEO number
+"defaultDbNames"
+
+#' Annotations of Motifs to Transcription Factors (TFs)
+#' 
+#' @rdname motifAnnotations
+#' @name motifAnnotations
+#' @aliases motifAnnotations_hgnc motifAnnotations_mgi motifAnnotations_dmel
+#' @title Annotations of Motifs to TFs
+#' @description 
+#' This dataset contains annotations linking motifs to transcription factors (TFs) 
+#' for different organisms (Human, Mouse, Fly) based on motif collection version 9.
+#' 
+#' \bold{Version 9}: 
+#' Annotations for motif collection version 9 ('mc9nr').
+#' \itemize{
+#'     \item{Human: motifAnnotations_hgnc_v9 (\emph{motifs-v9-nr.hgnc-m0.001-o0.0.tbl})}
+#'     \item{Mouse: motifAnnotations_mgi_v9 (\emph{motifs-v9-nr.mgi-m0.001-o0.0.tbl})}
+#'     \item{Fly: motifAnnotations_dmel_v9 (\emph{motifs-v9-nr.flybase-m0.001-o0.0.tbl})}
+#' }
+#' 
+#' These objects are designed for use with RcisTarget without modification, 
+#' but users can explore them to get more information about the motifs. 
+#' The original files are available from \code{https://resources.aertslab.org/cistarget/motif2tf/}.
+#' 
+#' \bold{Columns:}
+#' \itemize{
+#'  \item{\bold{motif}: }{Motif ID.}
+#'  \item{\bold{TF}: }{Transcription factor or inferred gene.}
+#'  \item{\bold{directAnnotation}, \bold{inferred_Orthology}, \bold{inferred_MotifSimil}: }{Boolean values indicating whether the motif is annotated to the TF 
+#'      directly ("directAnnotation") or inferred by orthology ("inferred_Orthology") or motif similarity ("inferred_MotifSimil").}
+#'  \item{\bold{Description}: }{Description of the annotation source.}
+#'  \item{\bold{annotationSource}: }{Source of the annotation as a factor. Levels: directAnnotation, 
+#'      inferredBy_Orthology, inferredBy_MotifSimilarity, inferredBy_MotifSimilarity_n_Orthology.}
+#' }
+#' 
+#' @docType data
+#' @seealso importAnnotations, RcisTarget
+#' @keywords datasets
+NULL
+
+

R/runSCENIC.R

@@ -0,0 +1,85 @@
+#' wrapper function for running SCENIC on default parameter. 
+#' 
+#' https://htmlpreview.github.io/?https://github.com/aertslab/SCENIC/blob/master/inst/doc/SCENIC_Running.html
+#' 
+#' @param sce A singlecellExperiment object.
+#' @param species The species of the data, one of human, mouse or drosophila_melanogaster.
+#' @param nCores The number of cores to use for parallel processing.
+#' @param dbDir The directory path to the cisTarget databases files.
+#' @param assay_name the name of assay.
+#' 
+#' @return The class contains the options/settings for a run of SCENIC. 
+#' 
+#' @importFrom SummarizedExperiment assay assayNames
+#' @importFrom utils data
+#' @export 
+runSCENIC <- function(sce,species,nCores,dbDir,
+                      assay_name = "counts"){
+  # Create directories to store process data and output results
+  if (!dir.exists("int")) {
+    dir.create("int")
+  }
+  if (!dir.exists("output")) {
+    dir.create("output")
+  }
+
+  # Initialize settings
+  org <- switch(species, 
+                "human" = "hgnc", 
+                "mouse" = "mgi", 
+                "drosophila_melanogaster" = "dmel", 
+                stop("Unsupported species"))
+  data(defaultDbNames,envir = environment())
+  dbs <- SCENIC::defaultDbNames[[org]]
+
+  if (org == "mgi") {
+    data(list = "motifAnnotations_mgi",envir = environment())
+  } else if (org == "hgnc") {
+    data(list = "motifAnnotations_hgnc",envir = environment())
+  } else if (org == "dmel") {
+    data(list = "motifAnnotations_dmel",envir = environment())
+  } else {
+    stop("Unsupported species. Please choose 'mgi', 'hgnc', or 'dmel'.")
+  }
+
+  scenicOptions <- SCENIC::initializeScenic(org = org, dbDir = dbDir, dbs = dbs, nCores = nCores)
+
+  # Process the cell annotation information of the sce object for plot later
+  if ("CellType" %in% colnames(SingleCellExperiment::colData(sce))) {
+    cellInfo <- as.data.frame(SingleCellExperiment::colData(sce))
+    saveRDS(cellInfo, file="int/cellInfo.Rds") 
+    scenicOptions@inputDatasetInfo$cellInfo <- "int/cellInfo.Rds"
+} else {
+  warning("Warning: 'cellInfo' not found, proceeding without cell annotations.")
+}
+
+  if ("CellTypeColor" %in% colnames(SingleCellExperiment::colData(sce))) {
+    colVars <- as.list(sce$CellTypeColor)
+    saveRDS(colVars, file="int/colVars.Rds")
+    scenicOptions@inputDatasetInfo$colVars <- "int/colVars.Rds"
+} else {
+  warning("Warning: 'colVars' not found, proceeding without color annotations.")
+}
+
+  # Co-expression network 
+  exprMat <- SummarizedExperiment::assay(sce, assay_name)
+  genesKept <- SCENIC::geneFiltering(exprMat, scenicOptions = scenicOptions,
+                                     minCountsPerGene = 3 * 0.01 * ncol(exprMat),
+                                     minSamples = ncol(exprMat) * 0.01)
+  exprMat_filtered <- exprMat[genesKept, ]
+  SCENIC::runCorrelation(exprMat_filtered, scenicOptions)
+
+  set.seed(123)
+  exprMat_filtered <- log2(exprMat_filtered+1) 
+  SCENIC::runGenie3(exprMat_filtered, scenicOptions)
+
+  # Build and score the GRN
+  exprMat_log <- log2(exprMat+1)
+  scenicOptions <- SCENIC::runSCENIC_1_coexNetwork2modules(scenicOptions)
+  scenicOptions <- SCENIC::runSCENIC_2_createRegulons(scenicOptions)
+  scenicOptions <- SCENIC::runSCENIC_3_scoreCells(scenicOptions, exprMat_log)
+
+  saveRDS(scenicOptions,file="int/scenicOptions.Rds")
+  return(scenicOptions)
+}
+

inst/prototype/read_SCENIC_dataset.R

@@ -0,0 +1,17 @@
+## code to load the databases for runSCENIC
+
+# load the defaultDbNames.RData from SCENIC
+load("./inst/ext/defaultDbNames.RData")
+usethis::use_data(defaultDbNames, overwrite = TRUE)
+
+# load the motifAnootation of three species from RcisTarget
+load("./inst/ext/motifAnnotations_dmel_v9.RData")
+motifAnnotations_dmel <- motifAnnotations_dmel_v9
+load("./inst/ext/motifAnnotations_mgi_v9.RData")
+motifAnnotations_mgi <- motifAnnotations_mgi_v9
+load("./inst/ext/motifAnnotations_hgnc_v9.RData")
+motifAnnotations_hgnc <- motifAnnotations_hgnc_v9
+usethis::use_data(motifAnnotations_dmel, overwrite = TRUE)
+usethis::use_data(motifAnnotations_mgi, overwrite = TRUE)
+usethis::use_data(motifAnnotations_hgnc, overwrite = TRUE)
+