reasons for htseq-count with zero features

Question

[shenwei1376]

Hi everyone,

I have asked this before, but want specify more this time. I have been using STAR-Htseq analyze the RNASeq data. Everything was find until recently, one of the library gives me 0 counts after the htseq-count. All the other libraries are fine, which means the gtf file or pipeline is find. I am wondering what could be reasons for 0 zero counts?

Many thanks for any discussion!

Best,

Wei S

[Ram]

I think auto-correct changes a bunch of fines to finds. Can you edit your post and correct those please?

[h.mon]

Duplicate of htseq-count with zero features.

[Ram]

shenwei1376 , please do not open duplicate questions, especially when queries on the original are yet to be addressed. Doing so is rude to people helping you in the other post and does not make you look good.

Answer 1

[GenoMax]

Have you looked at the alignment report for that sample and if so what does it say. Perhaps you received data that is not from your sample or the sample may have failed during library prep and may need to be repeated. Either possibility will lead to 0 counts. Take a few reads and blast them at NCBI to discount first possibility. For the latter you will need to dig into QC and/or talk with your sequence provider.

Note: This is based on your assertion that rest of the samples have worked well and produce reasonable results with the pipeline you are using.

[shenwei1376]

Thank you for your suggestions!! The library has been mapped to two indexed genomes. One is indexed based on a gtf file that is re-structured, so all the genes are re-grouped into transcripts (protein informs). Another one is indexed based on regular gtf file. The strange thing is all libraries worked when mapped to first genome, but the last library failed when mapped to the second genome! The pipeline and settings are same!

Thank you for more thoughts!