Is primer removal necessary from reads? What if the primers be present in the reads’ ends? Is it problematic?
Sharing your mind will make me happy.
I think is good practice to get rid of any ‘artificial’ sequences. During the PCR step, the polymerase includes the primer molecules into the amplicons, so you loose the bacterial sequence (and with it any informative variation in it) which is replaced by the sequence you ordered.
Primer sequences usually contains degenerated bases, that means different amplicons obtained from the same 16S transcripts may include different primer variants. This may result in confusing the classifier during the taxonomy assigning steps.
If the primer sequence is present also in the end of the reads (3’ side, mostly depends on your experimental design I guess ), I suspect it may get in the way of the merging step in dada2 if not considered (by trimming or dada2 parameters).
Overall, I think trimming primers is not a difficult step (at least within the qiime2 pipeline) and make things much easier later, so it is not worthy to skip it!
This is my 2cent, hope it helps