I have a question about truncation parameters for a 2x300 bp paired end run using the V3-V4 region. These results are with the primers removed (341F and 805R). Here is the visualization of demux.qzv after trimming primers using cutadapt.
My first data point is with reads that have been truncated at --p-trunc-len-f 281 and -p–trunc-len-r 202. Theoretically, these limits should be fine given that the expected amplicon size is around ~460 bp. This should leave enough for the ~20 bp DADA2 overlap requirement. As you can see, not much is lost at merging suggesting that this step is working fine.
I followed up by truncating a bit deeper, to see if the reads still merge. These parameters were set at --p-trunc-len-f 245 and -p–trunc-len-r 191.
Now although I lose a little bit more reads at the merging steps compared to the first truncation, I still retain a good amount. However, if the amplicon size is around ~460 bp, shouldn’t I be losing a lot more reads at merging, especially accounting for the ~20bp DADA2 overlap requirement? What could explain retaining such a large amount of reads?