Ask Your Question

what does "intronic depth" mean in tophat alignment visual?

asked 2013-09-15 16:45:39 -0600

Bhakti Dwivedi gravatar image


I have generated tophat alignment from RNAseq data with default settings. When I visualize the alignment on Golden Helix genome browse, I notice these light grey stacked pile of reads labeled as "intronic depth". These seem to typically map over intronic regions and have a good coverage but I can't see the base or any other information related to it. For instance, I have a position in the mapped alignment happen to be in a coding gene exonic region marked with:

match G 2 reads mismatch A 20 reads intronic 150 reads

Does anyone know what does this "intronic depth" means? How can I explain these intronic reads in a coding region where I see a somatic change? are these of concern? should such site with a variant be considered as genuine?

I would appreciate any feedback or help. Thank you!


edit retag flag offensive close merge delete

1 answer

Sort by ยป oldest newest most voted

answered 2013-09-16 15:26:36 -0600

MichaelThiesen gravatar image

Good Question!

When sequencing RNA, you are usually looking at post-spliced RNA fragments. These reads don't contain any detailed information about intronic nucleotides because by the time they are sequenced, the intronic RNA has been spliced out. The only thing the aligner can tell us is that the read spans an intron. The intronic depth at a location is the number of reads that span the location, but don't have any sequenced RNA for that region.

So, looking at your example, it seems that 150 out of 172 reads simply skip over the exon in question. This probably means that the exon is poorly expressed and gets spliced out most transcripts. However, in the 22 cases where the exon wasn't spliced out, 20 had a mismatch. I'd say you have some good evidence for a variant, but I'm not sure it's conclusive. A common standard for calling a variant is to have at least 30 reads that support it.

edit flag offensive delete link more
Login/Signup to Answer

Questions should be tagged FeatureRequest for asking about a non-existing feature or proposing a new idea, GeneralInquiry for general questions about GenomeBrowse or directions on how to do something, or RanIntoProblem if you want to report an issue or had difficulty getting to an expected result.


Asked: 2013-09-15 16:45:39 -0600

Seen: 1,015 times

Last updated: Sep 16 '13