Failing to re-alligning sequences with same truncLen parameters

[Debora6991]

Hello,
I've been an intensive user of your dada2 pipeline until a few years ago. After 4 years, I need to re-run the same data in the pipeline to map the reads agains an updated database (2x300 paired end, V3-V4 region).
My samples present with these quality profiles after removing the primers with Cutadapt:

I've tried to fine-tune the truncLen parameters from 200 to 250 for forward reads and 160 to 190 for reverse reads in all possible combinations, but I always get this output (or very similar) after the merging step

The only combination that raises the merged reads a bit (15%) is 200 forward/170 reverse. Above and below these values, the merged reads drop again to 1.9%.

I am wondering what I am doing wrong, as a few years back I successfully merged reads with an average output abundance close to 50%. I compared the script used back then and now, and nothing significant has changed.

Could you please provide some guidance on how to resolve this issue? Thank you!