TrimGalore! not properly removing low quality base calls.

Question

[Bio_info_matics]

Hello all,

I've recently been attempting to trim a set of scRNA-Seq fastqs for alignment and analysis which show fairly poor quality using TrimGalore with the following parameters:

~/TrimGalore-0.6.6/trim_galore --path_to_cutadapt ~/.local/bin/cutadapt --illumina --paired --phred33  -q 20 -j 8 -o outdir R_1.fq.gz R_2.fq.gz

This successfully removes adapters, but does not appear to remove low quality base calls (<20Q) as I'd expect. I had used TrimGalore some time ago and didn't have this problem.

Help or explanation would be appreciated! Maybe I'm misinterpreting something.

[gb]

You can find the explanation here: https://cutadapt.readthedocs.io/en/stable/algorithms.html#quality-trimming-algorithm

Assuming your screenshot is based on trimmed data you can see that the quality around the 150 is higher than lets say position 120. I guess the quality of position 150 is high enough to make the trimming algorithm stop.