Drop ASC sites in blacklist / low-mappability regions

Repetitive and low-mappability regions are where WASP fails to remove reference mapping bias AND where GATK makes false-heterozygous genotype calls (reads from a paralog map in). Both push significant ASC toward the reference allele. Since we lack the paper's imputed-genotype QC, removing these regions is the standard, faithful mitigation (Calderon et al. also excluded blacklist regions in QC).

ascFilterRegions(snpIds, regionsGr, verbose = TRUE)

Arguments

snpIds

character vector of "chr:pos_REF_ALT" ids (rownames of the count matrices).

regionsGr

GRanges of regions to remove (ENCODE blacklist and/or low-mappability). Seqnames must be UCSC-style ("chr1", ...).

verbose

log how many sites were removed (default TRUE).

Value

the subset of snpIds NOT overlapping regionsGr.

Author

Irem B. GUNDUZ