Quality control ChiP-Seq

Question

[marina-orlova]

Hello

I'm trying to improve quality of reads (ChiP-Seq, drosophila melanogaster).

Initially the quality was very bad

I tried using Trimmomatic with parameters TRAILING:20 and MINLEN:50

After that, per base quality became good.

Next, I searched for overrepresented sequences with BLAST and found out that most of them are from drosophila genome, others (such as GAAGAGAAGA, TTTTTTTTT etc.) - are not from it. Besides, these overrepresented sequences have much in common.

Please advise what I should do with overrepresented sequences - cut them or don't do anything.

[Ming Tommy Tang]

If those are not adapters, keep the overpresented sequences, but trim off the bad quality bases by Trimmomatic

Answer 1

[Devon Ryan]

Don't do anything unless they're adapters, which they probably aren't since you already trimmed things.

BTW, depending on what exactly was being IP'd, one might expect some highly overrepresented sequences (in fact, that might be the whole point of the experiment).

[marina-orlova]

So, do you think I should just start mapping in spite of the low quality showed by fastqc?

[Devon Ryan]

You should always quality and adapter trim. Aside from that, just map the remaining reads.

[marina-orlova]

Thank you for the answer!