How to resize a GenomicRange objects centered on a DNA motif

Question

[Ming Tommy Tang]

Hi there,

I have a bed file from MACS, I did the motif-enrichment by MEME-ChIP and found the primary motif. Now, I want to find the secondary motif, but I want to resize the bed file to 500bp centered on the primary motif. Is there an easy way to do this? I know one can use resize function for GenomicRanges object, but it can only center on the middle of the ranges.

Thank you,

Ming

Answer 1

[seidel]

If you know the location of the primary motif for the features in your bed file, simply use the starts and ends of the motifs to create a new GRanges object, and resize that with a fixed center.

Answer 2

[micahgearhart]

This function returns an aligned GRanges object based on matchPattern or will also work with matchPWM. It's a little slow because it has to retrieve the DNA sequence for each range but I wanted to make sure there was only one site under each peak.

returnAlignedGR <- function(gr,pattern,width) {
#  pattern<-"WWWGWWGAANTTCWWCWWW"
  c <-as.matrix("chr0",nrow=length(gr))
  s <- as.matrix(rep(0,width(gr[1])))
  for (i in 1:length(gr)) {
  seq<-as.character(seqnames(gr[i])@values)
  c[i]<-seq
  start<-ranges(gr[i])@start
  width2<-ranges(gr[i])@width
  peak <- DNAString(Mmusculus[[seq]],start=start,nchar=width2)
  site <-  matchPattern(pattern,peak,fixed=F)
  if(length(site) == 1) {
                          s[i]<-start+start(site)+8 #assuming motif is 15mer, 8 is center
                          } else {
                            s[i]= 0
                          }
  }
  gr_outRD<-RangedData(space=c, IRanges(start=s-width, end=s+width ), strand="*")
  return(as(gr_outRD,"GRanges"))
  }