HTSeq result, most genes have 0 count?

Question

[bxia]

Hi, I use HTSeq to map my reads to annotation and output as gtf file, but I didn't see any genes with count more than 0...

What could be the potential problem?

[komal.rathi]

Well, I would ask any question with some background and basic information, for instance:

1. What was the reference used for alignment of your reads? What species and build?
2. What gtf annotation and build are you using?
3. Are your reads stranded or unstranded?
4. Are the reads single-end or paired-end?
5. What was the exact command used for htseq-count?
6. Maybe add what aligner you used to align your reads and what was the exact command used. It might not be relevant to this question but it never hurts to give a context of your problem.
7. Paste lines of relevant code as well as error(s)/warning(s) that you get.
8. Paste some lines of relevant files. See head.

[bxia]

Looks like most of reads not mapped uniquely, is that caused by the alignment?

I used Hisat2 to map the reads.

[WouterDeCoster]

Could you be a little more specific on this? What were the alignment stats reported by hisat2? What is in the bottom lines of the htseq output? (__no_feature and __ambigious, etc)

The more information you give us, the easier we can track down your problem.

[bxia]

__no_feature 18304240 __ambiguous 132877 __too_low_aQual 0 __not_aligned 4197952 __alignment_not_unique 7001070

[bxia]

The corresponding mapping results:

4197952 (16.42%) aligned 0 times 16784289 (60.57%) aligned exactly 1 time 5352503 (22.15%) aligned > 1 times

[GZ1995]

Check out whether your annotation gtf file and reference genome file are from the same source. Avoid using these files from different sources (e.g., USCS gtf file and Ensemble reference genome).

[bxia]

I download the annotation and genome index from igenomes

[GouthamAtla]

Following with your previous question, there could be a problem with annotation format. Could you post first 2 - 3 lines of your annotation file and the command you used ?