scip-workflows 0.1.1

No project description provided

SCIP use case workflows

This repository contains Snakemake workflows to reproduce use cases presented in A scalable, reproducible and open-source pipeline for morphologically profiling image cytometry data.

It is built using two frameworks:

• nbdev
• Snakemake

nbdev is a framework for developing reusable code in notebooks. Functions are defined and tested in notebooks, and exported to a package. This package can be installed and reused in other notebooks or scripts.

Snakemake is a workflow framework to create reproducible data analyses. Workflows are defined via a human-readable language, and can be easily executed in various environments.

Installation

To execute the workflows in this repository you need to install Snakemake.

The workflows have been tested with Python 3.9.13.

Usage

Reproducing the use cases is done by executing SCIP to profile the images, and Snakemake workflows to generate downstream analysis results.

The required configurations to run SCIP are in the scip_configs directory.

The following commands expect Snakemake to be available. Snakemake can be executed using conda environments or a pre-existing environment containing all required packages.

To reproduce a use-case, open a terminal where you cloned this repository and execute:

snakemake --configfile config/use_case.yaml --directory root_dir use_case

where

• use_case is one of WBC, CD7 or BBBC021,
• root_dir points to where you downloaded the use case files

Make sure to update the config file to your situation; mainly setting the parts to the amount of output partitions SCIP generated.

This expects the environment to contain all required dependencies. Add --use-conda to let Snakemake create a conda environment containing all requirements.

Use case: WBC

Data and features (for SCIP and IDEAS) can be downloaded at the Bioimage Archive

Use case: CD7

Data and SCIP features can be downloaded at the Bioimage Archive

Use case: BBBC021

Data can be downloaded at the Broad Bioimage Benchmark Collection. Features can be downloaded from Zenodo. You can download metadata BBBC021_v1_image.csv and BBBC021_v1_moa.csv from the supplementary materials "Data S2" in [1].

[1] Ljosa, V., Caie, P. D., Ter Horst, R., Sokolnicki, K. L., Jenkins, E. L., Daya, S., ... & Carpenter, A. E. (2013). Comparison of methods for image-based profiling of cellular morphological responses to small-molecule treatment. Journal of biomolecular screening, 18(10), 1321-1329.