Counting reads for gene transcripts with overlapping loci

Question

[Ruet]

I am working with the human genome and have been allowing single mapping during alignment with Tophat. When I count reads that map to a particular loci, if a read falls on two gene transcripts, I do not count the ambiguous read. The transcripts I am interested in are snRNAs and snoRNAs but is seems like the transcripts of interest are falling under longer transcripts. Therefore, I count reads from the smaller snRNA transcripts as well as some reads from the longer transcript. I was wondering how to counts reads for overlapping transcripts especially when the shorter gene transcripts falls entirely under a longer gene transcript?

[mark.ziemann]

Is your RNA-seq protocol stranded? Are you using this information in your read counting?

[Ruet]

Hi Mark,

Yes, my RNA-seq protocol is stranded and I am using the information for read counting in my analysis. I am using the "union" model in the figure below, and I am not counting ambiguous reads.

Thanks!