when using htseq-count, should I only count read-pairs whose two mates were both reported ?

Question

[Pei]

Hi all:

Working with pair-end RNA-seq data, I found that for some read-pairs, only one end/mate was reported by tophat-output sam file and then htseq-count. For such cases, what would you do in usual when calculating gene-wise read count?

(I think I should only count those read-pairs whose both ends were successfully aligned and reported by htseq-count.)

Your opinions must be very valuable.

Thanks in advance!

Best

Answer 1

[Devon Ryan]

I prefer to count orphans as well. Sometimes one of the reads just has crap quality and doesn't align, I see no reason to deflate counts because of that.

[Pei]

Thank you!!