View Single Post
Old 04-12-2013, 01:07 PM   #2
Julien Roux
Location: Chicago

Join Date: Dec 2011
Posts: 24

I am in the process of writing a perl script that generates the file junctions.bed from a given tophat alignment BAM file and I discovered a small bug that sometimes affects the definition of the flanking regions covered by mapped reads next to the exon-exon junction. When there is an insertion or a deletion on one side of the junction, the number of flanking covered nucleotides is not counted correctly.
For example in the following cigar line 14M1I90M664N2M, there are 14+90 nucleotides covered on the left of the junction, but tophat erroneously counts only 14. The same thing happens when there is a deletion.

I thought this would be of interest for people working on related matters.
Julien Roux is offline   Reply With Quote