Row-specific findOverlaps of two GRanges datasets

Question

[RogelioG]

Hi all,

I am working on a way to find sense/antisense encoded transcripts and the amount of overlap they have.

One final piece of code I cant get together: I have a two dataframes (GRanges) with EnsT ids and start and end position of each transcript, both for sense and antisene encoded transcript. From this I want to calculate the amount of overlap of the two corresponding sense/antisense transcripts that are in the same row.

Here comes the issue: findOverlaps searches all the data of a GRanges object, instead I need it to only search for overlap in the same row number for both GRanges object. Probably easy to do, but cant find a solution in the Documentation.

Thanks beforehand!

RG

[Kevin Blighe]

It would help to paste some sample data and to show the result that you want.

[RogelioG]

data and part of script I am running:

A <- makeGRangesFromDataFrame(a)
B <- makeGRangesFromDataFrame(b)

A             seqnames               ranges strand
                <Rle>            <IRanges>  <Rle>
12 ENSMUST00000138535 [61767414, 61851540]      *
21 ENSMUST00000016309 [74288246, 74304336]      *
25 ENSMUST00000027431 [86099036, 86111116]      *
26 ENSMUST00000027431 [86099036, 86111970]      *
28 ENSMUST00000113212 [87384576, 87394729]      *

B            seqnames               ranges strand
                <Rle>            <IRanges>  <Rle>
12 ENSMUST00000129030 [61638823, 61769859]      *
21 ENSMUST00000087226 [74285033, 74353692]      *
25 ENSMUST00000080204 [85600702, 85645036]      *
26 ENSMUST00000054279 [85649987, 85683118]      *
28 ENSMUST00000152501 [86021941, 86024669]      *

... ... ... .

olaps <- findOverlaps(A, B, select="all",type="equal",use.region="first")

isect <- pintersect(A[queryHits(olaps)], B[subjectHits(olaps)])

olaps2 <- data.frame(query=queryHits(olaps), subject=subjectHits(olaps),olap_width=width(isect))

[Kevin Blighe]

Thanks! - if I just run findOverlaps with the default, I believe that it does what you need, no?

olaps <- findOverlaps(A, B)

isect <- pintersect(A[queryHits(olaps)], B[subjectHits(olaps)])
isect
GRanges object with 2 ranges and 1 metadata column:
      seqnames               ranges strand |       hit
         <Rle>            <IRanges>  <Rle> | <logical>
  [1]       12 [61767414, 61769859]      * |      TRUE
  [2]       21 [74288246, 74304336]      * |      TRUE
  -------
  seqinfo: 5 sequences from an unspecified genome; no seqlengths


olaps2 <- data.frame(query=queryHits(olaps), subject=subjectHits(olaps), olap_width=width(isect))
olaps2
  query subject olap_width
1     1       1       2446
2     2       2      16091

If It's not quite what you need, then just do some extra filtering on the olaps2 object, like:

olaps2[,olaps2$query==olaps2$subject]

This will show just the overlaps on the same row