FastQC per base sequence content

Question

[Thanh]

I'm running FastQC on some paired-end fastq files. I have a warning on per-base sequence content, as the first 5 to 6 bases show significant bias towards T and G, as shown below. I was wondering what the sequence in the first 5 or 6 bases is and whether I should trim it out before aligning with bwa.

Answer 1

[GenoMax]

What kind of data is this? You can see this blog (likely applicable in your case): https://sequencing.qcfail.com/articles/positional-sequence-bias-in-random-primed-libraries/

[Thanh]

This is WES data with paired-end reads. Thank you for the reading! So I gather from this article that trimming the 5' end would be unnecessary for the downstream analysis, right?

[GenoMax]

Correct. You can verify that by alignment stats.