Fast Rna-Seq Exon Count Calculation

Question

[lkmklsmn]

Hi,
I want to calculate exon counts from a large number of bam files. I have done this before using HTseq. Since I have a large number bam files I really need to make this process as efficient as possible. Which programs would you recommend?
I am using cloud resources with multiple CPUs.

Answer 1

[Devon Ryan]

Just invoke multiple instances of htseq-count, one per file. I routinely count all of my samples at once that way. Allegedly featureCounts from subread is faster, but I've not used it.

[Mikael Huss]

featureCounts is really fast. It usually takes less than a minute per sample (which obviously depends on the size of the sample). That said, I still use HTSeq out of habit because it's fast enough.

[Devon Ryan]

Good to hear from someone who's used featureCounts (and isn't its author) that it really is that fast! Was it relatively straight-forward to get working?

[Mikael Huss]

Yes, I had no problems getting it to run. I have been using it in R through the Rsubread package, pretty much following the manual.

[Sean Davis]

I'll second the sentiment that featureCounts is a nice speed increase.