Calculate Read Counts Rnaseq By Htseq-Count

Question

[camelbbs]

Hi, I want to ask if HTSeq-count will over-counts the reads in overlapped exons?

IF I give the refseq.gtf as gtf annotation, there should be some isoforms overlap exons, so how htseq-count handle this.

I think htseq-count will give counts for each isoform separately.

And I calculate rpkm for each isoform by myself.

And then, for one gene have many isoforms, I calculate the average rpkm for this gene. Is this procedure OK?

[Eric Normandeau]

Hi Camelbbs. I'm adding this comment to all your questions: Please take some time, before you ask a question, to think more about your problems and most likely sources of answers (manuals, FAQs, Google!, etc.). When you ask a question, include some context, tell us why you ask that question, what result you need, etc. Most of your questions are vague, impossible to answer or you changed them following an answer because it became evident that it was not clear. Cheers.

Answer 1

[Steve Lianoglou]

Hi, I want to ask if HTSeq-count will over-counts the reads in overlapped exons?

IF I give the refseq.gtf as gtf annotation, there should be some isoforms overlap exons, so how htseq-count handle this.

I'm pretty sure that the documentation, specifically the section about counting, addresses this specifically. If that page does not address your question, it would be helpful if you could modify your question to be a bit more specific about what the confusion is.

I'm sorry, I do not follow what your question about rpkm is. Are you suggesting that you want to calculate the RPKM for each isoform of a gene individually and then take the average of this number? Are you proposing to use only reads that can unambiguously be assigned to one isoform for this?

You ask if it is OK, but you do not mention what the purpose of calculating expression in this way is for, which makes it impossible to judge.