I understand that the contig pairing algorithm in mothur (make.contigs) does not work very well with MiSeq V3V4 data (with the typically poor R2 quality). I was told that USEARCH does this part better. Is there any reason not to use USEARCH for contig pairing, then continuing along the mothur MiSeq SOP?
The problem lies in the fact that by sequencing the V3-V4 region you do not get full overlap of the forward and reverse reads, and therefore do not get the ability to correct low quality bases with the higher quality bases in the complementing strand. The problem does not lie in what software is used to do this step, so using USEARCH for contig pairing will not overcome this issue.