Question

Extract coordinates of overlapping genomic intervals

1

Entering edit mode

rubi ▴ 110

@rubi-6462

Last seen 5.7 years ago

Hi,

I have two sets of Genomic Ranges which I'm intersecting using the findOverlaps of the GenomicRanges package:

df1 <- data.frame(chr=rep("chr1",6), start=c(10033259,10060726,98674166,10067579,10067607,11169988), end=c(10033289,10060783,98674223,10067654,10067664,11170044), strand=c("-","-","+","+","+","+"))

df2 <- data.frame(chr=rep("chr1",3),start=c(10024601,10033258,10033258),end=c(10038168,10033323,10033323),strand=c("-","-","-"))

df1.gr <- makeGRangesFromDataFrame(df1,seqnames.field="chr",start.field="start",end.field="end",strand.field="strand")

df2.gr <- makeGRangesFromDataFrame(df2,seqnames.field="chr",start.field="start",end.field="end",strand.field="strand") dfs.ol <- findOverlapsdf1.gr,df2.gr)

My question is how to extract the actual overlapping coordinates of each of the hits in the returned value of findOverlaps (dfs.ol)?

I know that the intersect function returns the collapsed intervals in the query genomic ranges which intersect with a search genomic ranges. But what I really need for each overlap between gr1 and gr2 are the coordinates of the overlap, in addition to the indices of the genomic ranges which overlap (in the returned Hits object).

genomicranges findoverlaps intersect • 1.6k views

ADD COMMENT • link updated 7.7 years ago by Steve Lianoglou ★ 13k • written 7.7 years ago by rubi ▴ 110

score 2 · Accepted Answer · 2016-08-08

2

Entering edit mode

Jeff Johnston ▴ 90

@jeff-johnston-6497

Last seen 6.3 years ago

United States

You can use pintersect:

overlaps.gr <- pintersect(df1.gr[queryHits(dfs.ol)], df2.gr[subjectHits(dfs.ol)])

If you want all the results in one object, you can add the indices as metadata columns:

overlaps.gr$df1_hit <- queryHits(dfs.ol)
overlaps.gr$df2_hit <- subjectHits(dfs.ol)

ADD COMMENT • link 7.7 years ago Jeff Johnston ▴ 90

0

Entering edit mode

Does that report the overlap interval though?

ADD REPLY • link 7.7 years ago rubi ▴ 110

0

Entering edit mode

Yes, it generates the overlapping interval for each row (a query/subject pair) in your Hits object.

ADD REPLY • link 7.7 years ago Jeff Johnston ▴ 90

0

Entering edit mode

In the overlaps.gr object?

I only see the indices of the query and hit but not the overlap's coordinates. How do you extract that?

ADD REPLY • link 7.7 years ago rubi ▴ 110

0

Entering edit mode

As the overlaps.gr is a GRanges object, you can use start(), end() and seqnames() to extract the coordinates of the overlapping intervals.

ADD REPLY • link 7.7 years ago Jeff Johnston ▴ 90