Not really an error but when you use bin.seqs and then actually look at the pairwise distances included you see that the cutoff is made at the 0.0X level… Which is also indicated in the resulting file names, but if then I want to use this OTU grouping for a pub it is typically reported as 99 or 97% identity or whatever the cutoff maybe. The issue is that 0.01 cutoff for distance actually translates to an 98% identity cutoff, since seqs between 0.01 and 0.02 difference also join in.
Just sayin’
See the last question at: http://www.mothur.org/wiki/Frequently_asked_questions#Aren.27t_the_.27unique.27_and_.270.00.27_distance_levels_the_same.3F
Ya, I saw that… Example of where this causes problems though:
I’m making some IxD plots comparing trends at different OTU cutoffs. I want to call my 1st OTU 99% or greater identity. I use bin.seqs, my plot now has points just above 1%–major hassle to fix my matrices by hand.
I agree there is error in everything, and that OTUs are not necessarily very biologically meaningful, but I don’t see how introducing more error addresses this issue.
IxD?
It’s not adding more error, its accounting for it. If you don’t like it you can always set the precision to 1000 or 10000 and grab the cutoffs you want.