celigo_pipeline_automation_paper

celigo_pipeline_automation_paper

• Overview
• Documentation
• System Requirements
• Installation Guide
• Demos
• Example Outputs
• License

Overview

celigo_pipeline_automation_paper is a companion package for Automated hiPSC culture and sample preparation for 3D live cell microscopy. It contains algorithms to run an automated image-based colony segmentation and feature extraction pipeline to predict cell count and select wells with consistent morphology for high resolution 3D microscopy.

Documentation

For full package documentation please visit aics-int.github.io/celigo_pipeline_automation_paper.

System Requirements

Hardware requirements

celigo_pipeline_automation_paper package requires only a standard computer with enough RAM to support the in-memory operations.

Memory Requirements:

• Six Well: 64 GB (~ 20 min runtime)
• Ninety Six Well: 16 GB (~ 1 min runtime)

*Note: This process was originally intended for use in a High Processing Cluster. Running single files is possible with the given memory requirements, however, running tandem processes on a local computer is not reccomended.

Software Requirements

OS Requirements

This package is supported for Linux and macOS. The package has been tested on the following systems:

• Linux: Ubuntu 18.04.6 LTS
• macOS Monterey 12.5.1

Python Dependencies

celigo_pipeline_automation_paper mainly depends on the Python scientific stack.

Jinja2 ~= 3.1.2
Cellprofiler ~= 4.2.4

Installation Guide

There are a few components to this installation guide:

• Initial Setup
• Package Installation
• Ilatik Setup

Initial Setup

1. Install python <=3.9, Python Install

2. Make sure git is installed, How to install git.

   • Notes:
     • This package uses Github LFS to manage larger files, cloning without LFS installed will result in a full download of the testset.
     • If your computer uses msys to configure git, your base filename length is set to 260 characters. run:git config --system core.longpaths true to set this to default.

3. Clone Repository, How to clone a repository. Note: This Process can take 5-10 minutes for a standard computer.

4. Install make, celigo_pipeline_automation_paper uses a Makefile to organize its dependencies and development tools. How to install make.

Package Installation

Once the Initial Setup is complete, navigate to your cloned repository from the command line. Next, you need to build the virtual environment used for running this package. This is done using the Makefile and can be done with the following commands:

basic usage: `make install`
development usage: make `make install-dev`

Note: This package is dependent on Cellprofiler 4.2.4, which can be a difficult package to install given its many dependencies. We recommend testing the installation of the Cellprofiler package on your computer prior to installing this package. Cellprofiler Repo, Cellprofiler Package

Ilastik Setup

1. Install the Ilastik app to your local computer. This can be done from their website. Once you have done this, navigate to the Ilastik app directory and locate the run_ilastik.sh script.

2. Next you will need to manually edit the following in resource_paths.json:

   • "6_well_ilp" : [Path to 6 Well ILP],
   • "96_well_ilp" : [Path to 96 Well ILP],
   • "run_ilastik" : [Path to run_ilastik.sh]

The ILP files can be found under the pipelines folder, but are not specifically part of the package. After cloning this repository you can choose to move them to an external location or leave them in place.

Once this process is completed you will want to rerun make install or make install-dev:

make install 
make install-dev

Post-Processing

The post-processing operation performs a well-defined QC (Quality Control) morphology and imaging criteria-based scoring system, as described in this Nature Methods paper, following the execution of the cellprofiler pipeline. This process can be performed using a series of scripts in this repository. When running multiple wells or plates, each image's metrics are exported to subsequent CSV. We found it helpful to concatenate metrics using this Merge Script.

96 Well QC Scoring

• 96 well QC scoring is performed with: 96 Well QC Scoring

6 Well Confluency and Estimated Cell Count

• 96 well confluency and estimated cell count calculation is performed with: 6 Well Confluency and ECC

Hamilton Methods: Venus Method Pkg files

• Venus Methods files in the pkg format for the Semi Automated Passaging, Seeding and Maintenance, of the Allen cell collection of hiPSC lines expressing green fluorescent protein tagged to protein identifying specific cellular organelles and structures.
• You will find the instructions for the import in the word document in the folder. The method files should be compatible with the latest version of Venus and associated drivers. If you need assistance when importing and opening due to needed driver and/or library please contact Ben Gregor at [email protected] or your local Hamilton applications specialist.

Demos

Programmatic Example

from celligo_pipeline_automation_paper import CeligoSingleImage

image = CeligoSingleImage(raw_image_path = [RAW_IMAGE_PATH], working_dir = [WORKING_DIRECTORY])

downsample_output_path = image.downsample()
ilastik_output_path = image.run_ilastik()
cellprofiler_output_paths = image.run_cellprofiler()

CLI Example

The pipeline can also be run from the command line.

Single Image

celigo_pipeline_cli celigo_pipeline_cli --image_path "[RAW_IMAGE_PATH]" --working_dir "[WORKING_DIRECTORY]"

Directory

run_dir_cli  --dir_path "[RAW_IMAGE_PATH]" --working_dir "[WORKING_DIRECTORY]" --chunk_size [CHUNK_SIZE]

Example Outputs

6 Well

• 6 Well testset
• 6 Well expected output

96 Well

• 96 Well testset
• 96 Well expected output

License

This project is covered under the Allen Institute Software License.