Exon density in uniform bins across bam files

Question

[kevin.stachelek]

I'd like to calculate the percent of bases that overlap exons in uniformly sized (n mB) windows across several bam files. Could anyone help me determine this preferably using bedtools or R?

Answer 1

[kevin.stachelek]

using R GenomicRanges and given two granges objects exonic_bed and bins

hits <- GenomicRanges::findOverlaps(query = exonic_bed, 
subject = bins, ignore.strand = TRUE)
overlaps <- pintersect(bins[subjectHits(hits)], 
exonic_bed[queryHits(hits)],  ignore.strand = TRUE)

subjecthits <- bins[subjectHits(hits)]
subjecthits$n_overlap <- width(overlaps)

subjecthits <- 
  subjecthits %>% 
  as_tibble() %>% 
  dplyr::group_by(seqnames, start, end, width) %>% 
  dplyr::summarize(overlap = sum(n_overlap)) %>% 
  dplyr::mutate(percent_overlap = overlap/width)

[kevin.stachelek]

this seems relevant