Snakemake workflow for testing μ-PBWT against Durbin's PBWT and Syllable-PBWT on 1000 Genome Project (1KGP) phase 3 data. 1KGP data are available at this link.

Snakemake need to be already installed, for example via conda:

conda create -c conda-forge -c bioconda -n snakemake snakemake

cd muPBWT-1KGP-workflow
snakemake --cores <num_cores> --use-conda --resources load=100

The option --resources load=100 option will avoid using too much RAM for Durbin Algorithm 5 (about 500gb is still needed), running only one job at a time for the rule runPbwtIndexed(thanks Jan Schreiber).

The pipeline will generatecd some results:

• in results/data some useful CSV files
• in results/plots some plots in PDF format
• in results/tables some tables in LaTeX syntax