RNAseq: should I trim both 3' and 5'?

Question

[fr]

I'm running an RNAseq analysis that was sequenced in Illumina HiSeq 2500.

From the reference docs I can see that adapters are 10 bases long, so I guess that's the specification that I should give to Trim Galore.

I'm wondering if I should trim both 3' and 5' ends or just 5' ends? Should I consider the same 10 bases for both if applicable?

Thanks a lot

[GenoMax]

In Illumina sequencing adapter contamination should never be on 5'-end of the read unless you have plain primer dimers with no inserts or you are doing something unusual/custom.

Also take a look at this blog post from authors of FastQC.This is what you will likely find next.

[fr]

Thanks, found this thanks to your comment: https://emea.support.illumina.com/bulletins/2016/04/adapter-trimming-why-are-adapter-sequences-trimmed-from-only-the--ends-of-reads.html

Answer 1

[swbarnes2]

Don't just trim off 20 bases from the end like that. That's stupid, most reads don't have problems with adapters.

For starters, you can use fastqc to see if adapters are a real problem for your samples (ignore the other things fastqc tells you, most of them are not relevant to RNASeq)

If you have an adapter problem, use a trimming program to just remove the adapter sequence. You know that bcl2fastq can do this, so your fastqs might already be trimmed.