Hi just noticed that the Schloss SOP only includes 1 of the 2 steps in the AmpliconNoise processing (PyroNoise, but not SeqNoise). Is there a reason that shhh.seqs (SeqNoise) is not run in the SOP?
To expand the question, is the SOP using preclustering in place of SeqNoise? ie. do you recommend including both PyroNoise AND preclustering, rather than viewing these as alternative strategies?
Right. If you look at our PLoS ONE paper we do PyroNoise and then pre.cluster instead of SeqNoise/shhh.seqs. As we mention in the paper, there are too many parameters for our tastes in SeqNoise to appropriately tune. The problem we reported in the paper was that you were never really sure what the maximum number of differences would be between the ideal sequence and those that got recruited to it. With pre-cluster, this is the only parameter - diffs.
Pat