problem with pre.cluster topdown option


I get different results with pre.cluster(…, topdown=T) and pre.cluster(…,topdown=F). The latter removes much less sequences than the former does. I look at the code, and find that in the process() function, with topdown=F, there is a line to jump out of search when the first one is found. (j+=numSeqs; //exit search, we merged this one in.)
Isn’t this option just a way to present results?

Thanks for your help.

No. topdown=T starts with the most abundant sequence and looks for rarer sequences that are within ndiffs. topdow=F starts with the rarest sequences and looks towards the successively more abundant sequences. It was an idea we had that we never got anywhere with. Just use the default.