Variant calling

Some different workflows related to variant calling from existing genome alignments.

Testing

Running with snakemake and polars

Example of snakemake_env.yaml

name: snake8
channels:
  - conda-forge
  - bioconda
  - nodefaults
dependencies:
  - snakemake
  - polars

We can then create and activate this conda environment, and test the overall workflow should run with

conda env create -f snakemake_env.yaml
conda activate snake8
cd .test
snakemake -s ../Snakefile --configfile config.yaml -n

This should print out that jobs would be run, covering most pathways used in variant calling. The snakemake rules were designed to run using our snakemake profiles for Euler.

Usage

Many of the individual tools are assumed to be installed on $PATH, and so will almost certainly fail out of the box. This is a work in progress to be more reproducible.

Detailed pipeline steps

Config

Example snakemake command

Assuming you have the Euler profiles installed and are using snakemake v8+ with the slurm executor plugin, we can run

snakemake -s <path>/variant_calling/Snakefile --configfile config.yaml --profile "slurm/v8" --executor slurm --dry-run

Roadmap

Several ideas for improvement. Feel free to make pull requests with other solutions.

• containerize almost every rule
• add more general approach for imputation/filtering
• re-integrate mendelian analyses