tophat output

[zhaowei]

From Tophat output "junctions.bed", we could know the position of junctions. I think this junction should be intron. But after I assembled tophat output by Cufflinks, I found that the position of intron is different from that of intron.

[scottyler89]

A couple of possibilities:
Have you done QC filtering using something like FASTX toolkit? Or maybe, it's just optional exons, or an alternative splice site. Have you looked at the DNA sequence to see if it has possible splice acceptor/donor sites?

[zhaowei]

A couple of possibilities:
Have you done QC filtering using something like FASTX toolkit? Or maybe, it's just optional exons, or an alternative splice site. Have you looked at the DNA sequence to see if it has possible splice acceptor/donor sites?

I mean, if junction area is intron area, tophat output and Cufflinks output should be consistent. But when I checked a gene, the intron area implied by Tophat and Cufflinks are different. Then I noticed that many genes have this problem.

[scottyler89]

It might be worth double checking that the GTF file your using matches the bowtie alignment index. It's possible if they are from different builds of the same genome, the indexes for the alignments wouldn't match up with the expected intron/exon junctions.