Mapping paired-end reads, include also the SE-reads after trimming?

[JonB]

Hi,

I have trimmed paired-end reads with Trimmomatic, and some reads are therefore left as only single-end (one pair filtered out entirely). I am using tophat2 to map against a reference genome and my goal is differential gene expression analysis.

My question is if I should include also the single-end reads in the mapping? I am thinking that more reads are better, but are there any drawbacks of including both PE and SE reads?

Thanks!

Jon

[Brian Bushnell]

I am thinking that more reads are better, but are there any drawbacks of including both PE and SE reads?

More reads are better! The only drawback is that it makes the workflow slightly more complicated; depending on the aligner, you will generally need to map the paired reads and the singletons in two different passes, then merge the sam files. But I think the BWA can handle both in one pass by examining the read names.

[JonB]

Thanks! I think also Tophat can take a mix of paire-end and single-end reads.