Hello everyone, I have really good quality sequences (shotgun sequencing Novaseq6000) , but I want to filter and trimm them for MAGs assembly. I am using trimmomatic, and I cant achieve that more of 80% of sequences to pass the filter. How is that possible when they are long enough and with high Q values?

Am I using some parameters wrong?

I know LEADING/TRAILING and SLIDINGWINDOW could seem redundant but I am not being too strict anyway.

java -jar ~/miniconda3/pkgs/tr

immomatic-0.39-1/share/trimmomatic-0.39-1/trimmomatic.jar PE -phred33 UCT_350.raw.fq1.gz UCT_350.raw.fq2.gz output_forward_paired.fq1.g

z output_forward_unpaired.fq1.gz output_reverse_paired.fq2.gz output_reverse_unpaired.fq2.gz ILLUMINACLIP:/home/lemm/miniconda3/pkgs/tr

immomatic-0.39-1/share/trimmomatic-0.39-1/adapters/TruSeq3-PE-2.fa:2:30:10 HEADCROP:10 LEADING:20 TRAILING:20 SLIDINGWINDOW:4:30 MINLEN:80

Ideally I would like to preserve sequences with Q >= 32, length >=120.

Results here:

% Both surviving HEADCROP LEADING TRAILING SLIDINGWINDOW MINLEN

79.47 10 5 5 4:25 80

79.50 10 20 20 4:25 80

79.50 10 20 20 4:25 80

81.94 10 20 20 4:25 70

67.50 10 20 20 4:30 70

74.25 10 20 20 4:25 100

81.95 10 30 30 4:25 70

99.78 10 30 30 - 70

67.45 10 - - 4:30 70

It seems that the Slidingwindow is trimming too much, is that normal?

I´ve added images of the FastQC report so you can see.

Thank you!