Under data/AF_filter/variant_matrix/annotated_vars/{gene}_var_final.tsv I have list of variants that (CHROM, POS, ALT, REF, selected INFO etc). Under data/AF_filter/variant_matrix/gmats/{gene}_gmat.tsv I have a variant matrix (variant per row, and samples as columns where each value in the matrix is 0-2 or missing.
The total workfolow looks like this:
In rule get_gnomad in src/filtering.smk I append tag AF (should I use AF_popmax instead to match better?) I add variant frequency to the Variant list (output run_folder/annotated/{gene}_burden_gnomad.tsv) .
The filter_gene filteres genes based on some critera. The var_df.Distance <= params.distance_max is just filtering of distance to closest exon (in gene) and something I get from the creation of the genotype matrix (to match WES better).
Previous in the creation of the multisample VCF that we use (input for creation of genotype matrices) we already filtered on callrate < 95%, therefore I didn't include the 90% callrate filtration step.
Under src/FisherTest.R I rewrote the test to better match my input format. Hopefully, it should match your implementation.
Cheerio, Joel
The total workflow looks like this:
