I have paired end Illumina (NextSeq) data (Staph aureus, avg sequence length 150bp). After running FastQC, I found reads from Lane 4 R2 to have reduced quality compared to that in Lane 1-3, R1 and R2. With most Lane 4, R2 from my samples (100+ samples) I would need to trim 30 bps whereas I wouldn't need to trim more than 5 bps from the other lanes. Should I just discard the data from this lane? or trim the reads from Lane 4 R2 (and some R1) separately before concatenating the forward (and reverse).