CytoSnake: Orchestrating reproducible pipelines for processing high-dimensional systems morphology data with snakemake

Erik Serrano, Gregory P. Way University of Colorado Anschutz School of Medicine

Table of contents

• Table of contents
• About
• Installation
• Workflows

About

CytoSnake is a command line interface (CLI) tool that orchestrates reproducible workflows that process high-dimensional single-cell morphology features extracted from microscopy images. CytoSnake's workflows are written in Snakemake, which is a well established workflow manager that facilitates data reproducibility, scalability, and modularity.

CytoSnake makes it easy for user to process high-dimensional cell morphology data as it requires straightforward inputs and parameters. Below is an example on how to execute CytoSnake once installed:

```bash

setting up directory

cytosnake init -d -m -b

executing workflow

cytosnake run ```

note: -b is optional, it is used if there are multiple platemap files

Installation

Install CytoSnake with conda.

bash conda install -c bioconda cytosnake

NOTE: In case you don't currently have conda installed on your system, you can access the documentation here. We recommend using Miniconda, primarily due to its lightweight installation process.

To check if CytoSnake has been successfully installed, simply type cytosnake help to see the CLI documentation:

bash cytosnake help

Workflows

CytoSnake workflows are the main instructions on how your data is going to be processed. Each workflow comes with its appropriate configuration file.

Here is an example below:

```yaml annotateconfigs: params: inputdata: platedata joinon: - Metadatawellposition - ImageMetadataWell addmetadataidtoplatemap: True formatbroadcmap: False cleancellprofiler: True externalmetadata: "none" externaljoinleft: "none" externaljoinright: "none" compressionoptions: method: "gzip" mtime: 1 floatformat: null cmap_args: {}

aggregateconfigs: params: inputdata: annotated strata: - MetadataPlate - MetadataWell features: infer operation: median outputfile: none computeobjectcount: False objectfeature: MetadataObjectNumber subsetdatadf: none compressionoptions: method: gzip mtime: 1 float_format: null

```

Above is a portion of the listed configs from the cp_process workflow. Each block represents an analytical specific step that is conducted within the workflow. In this example, annotate_configs and aggregate_configs are separate steps that occur within the cp_process workflow. Each block has the params parameter, which are the parameters associated with the analytical step. Users can edit these parameters from the defaults if they want their workflow to analyze their data in a specific way.

Overall, each workflow will have a designated workflow config file. It will contain all the steps conducted in the workflow, and users have the option to change the default parameters that are specific to their dataset.