janis-pipelines.pipelines 0.9.1

Bioinformatics tools for Janis; the Pipeline creation helper

Janis - Example Pipelines

This repository contains workflows written using Janis.

It can be installed through PIP with:

pip3 install janis-pipelines.pipelines

This allows Janis to find these installed pipelines from the CLI (through entrypoint: janis.extension=pipelines) and means they can be imported from any Python file.

from janis_pipelines import WGSGermlineGATK

WGSGermlineGATK().translate("wdl")

Pipelines

Whole genome sequencing (WGS) pipelines:

• Germline
  • Germline (GATK only)
• Somatic
  • Somatic (GATK only)

Other:

• Alignment

Quickstart

To use these pipelines, you will need to have Janis installed:

pip3 install janis-pipelines

Then you can run the pipeline through Janis:

WORKFLOW="https://raw.githubusercontent.com/PMCC-BioinformaticsCore/janis-pipelines/master/workflows/alignment/alignment.py"

# Generate inputs file
janis inputs WGSGermlineGATK > myinps.yml

# Run workflow
janis run --inputs myinps.yml WGSGermlineGATK

WGS Germline pipeline

See the - germline folder for more information, the workflow and CWL / WDL translations.

The WGS germline pipeline takes a FASTQ pair, aligns, sorts, marks duplicates and calls variants across GATK4, Strelka and VarDict. These variants are combined and sorted at the end.

These variants were validated against the Genome in a Bottle data sets to achieve:

• Recall: 99.25%
• Precision: 92.02%

These results were identical across 3 research institutes (combination of Slurm / PBS / Torque) and Google Cloud platform. The pipeline took approximately 27-30 hours to run at a 30x coverage, depending on the resource constraint.

WGS Somatic pipeline for tumor-normal variant discovery

See the somatic folder for more information, the workflow and CWL / WDL translations.

The WGS somatic pipeline takes normal and tumor FASTQ pairs, aligns, sorts and marks duplicates separately, and then performs tumor-normal variant discovery across GATK4, Strelka and VarDict (in somatic modes). These variants are combined and sorted at the end.

This pipeline was run successfully across a similar set of research institutes (as germline) and the cloud, however these variants have not been validated yet.