The powerful awk script to calculate π, Dxy (or πxy, or Nei's D) and some more simple stats (Fst, Tajima's D, Ronfort's rho) in VCF files in the command line. Developed to analyze arbitrary-ploidy groups with substantial amounts of missing data.
⚠️ piawkais under development. If something does not seem to work well, check for updates and do not hesitate to file an issue!
conda install -c bioconda piawkaMake the following programs available in the command line (install and add to PATH):
gawkv5.0.0 and abovetabix
Then, get piawka by cloning the repo and add the scripts to PATH:
git clone https://github.com/novikovalab/piawka.git
export PATH="$( realpath ./piawka/scripts ):${PATH}"$ piawka
piawka v0.8.10
Usage:
piawka -g groups_tsv -v vcf_gz [OPTIONS]
Options:
-1, --persite output values for each site
-a, --all output more cols: numerator, denominator, nGeno, nMiss
-b, --bed <arg> BED file with regions to be analyzed
-B, --targets <arg> BED file with targets (faster for numerous small regions)
-D, --nodxy do not output Dxy
-f, --fst output Hudson Fst
-F, --fstwc output Weir and Cockerham Fst instead
-g, --groups <arg> either 2-columns sample / group table or
keywords "unite" (1 group) or "divide" (n_samples groups)
-h, --help show this help message
-H, --het output only per-sample pi = heterozygosity
-j, --jobs <arg> number of parallel jobs to run
-m, --mult use multiallelic sites
-M, --miss <arg> max share of missing GT per group at site, 0.0-1.0
-P, --nopi do not output pi
-q, --quiet do not output progress and warning messages
-r, --rho output Ronfort's rho
-t, --tajimalike output TajimaD-like stat (manages missing data but untested)
-T, --tajima output classic TajimaD instead (affected by missing data)
-v, --vcf <arg> gzipped and tabixed VCF fileSee the wiki for further details.
If you want to express your gratitude for having piawka, please cite our Siberian Arabidopsis paper where we have introduced and first used it.