How can I use featurecounts after generating a bam file using BWA?

Question

[a.rex]

I have a reference transcriptome, to which I have aligned reads using BWA. My goal is to quantify the aligned transcripts - how can I do this? Can I use featurescounts - bearing in mind that I have aligned to a transcript fa file.

[Ram]

I read yesterday about Salmon and Kallisto, which can generate count metrics from a transcripts.fa file and the FQ reads. Maybe they have an option to use an aligned BAM?

[ATpoint]

Salmon does, see the section of the manual :)

Answer 1

[h.mon]

If you aligned to the transcriptome, you should not use featureCounts, particularly if you have a well annotated transcriptome, or assembled the transcriptome yourself. In both cases, the transcriptome should have a lot of isoforms, which would result in a lot of multi-mapped reads, which would be discarded by featureCounts. See Why you should use alignment-independent quantification for RNA-Seq for a discussion on this.

Use RSEM in this case, or as ATpoint noted above, salmon quant.