RSEM HTSeq unique reads

Question

[riccardo]

Hi all, I have to build a STAR index with a human gene in the mouse genome and I want only the reads that map uniquely in these and other genes. Since I want also to distinguish the reads that map in these homolog genes in this case HTSeq is better than RSEM since it considers only the reads that map in a unique position? Thank you.

Answer 1

[h.mon]

No need for HTseq, STAR can handle both situations simultaneously:

--quantMode GeneCounts

• A read is counted if it overlaps (1nt or more) one and only one gene.

And:

--quantMode TranscriptomeSAM

• Output alignments translated into transcript coordinates, which can be used with RSEM.

To have both:

--quantMode GeneCounts TranscriptomeSAM

I think RSEM will be better, as it will probably make better use of the ambiguous reads mapping to both mouse and human genes. If the genes are too similar, though, you may end up with no counts, or unreliable counts.