summarizeOverlaps on stranded paired end RNAseq data

0

Entering edit mode

vedran.franke • 0

@vedranfranke-7218

Last seen 6.6 years ago

Germany

Does summarizeOverlaps work with stranded paired-end RNAseq data?

Will it assign fragments to the proper strand?

Best regards,

Vedran

summarizeoverlaps • 1.9k views

ADD COMMENT • link 8.7 years ago vedran.franke • 0

0

Entering edit mode

As I understand it, if you set the arguments ignore.strand = FALSE and single.end = FALSE, the two reads in a pair will be treated as coming from the same strand (presumably the strand of the first read), and you will get a hit if one or both ends overlap a feature.

ADD REPLY • link 8.7 years ago Mike Smith ★ 6.6k

0

Entering edit mode

I have set ignore.strand=FALSE, single.end=FALSE, fragments=TRUE and in BamFileList, asMates=TRUE.

The problem is that when I create 2 mock transcripts that have same coordinates and differ only in the strand, I

get almost the same number of reads for both transcripts, even though the fragments come exclusively from one strand (which can be seen in the browser).

ADD REPLY • link 8.7 years ago vedran.franke • 0

0

Entering edit mode

Can you show us an example?

When fragments=TRUE, the bam is read in with readGAlignmentsList(). You can explore this by reading in your data then calling the count mode on the reads and your annotation. See ?Union and the documentation for the 'inter.feature' arg.

bf = BamFile(myfile, asMates=TRUE)

reads <- readGAlignmentsList(bf)

Union(annotation, reads, ignore.strand=FALSE, inter.feature=TRUE)

Valerie

ADD REPLY • link 8.7 years ago Valerie Obenchain ★ 6.8k

Login before adding your answer.