Clustering at higher precision


I hope this hasn’t been addressed before. I would like to perform clustering at a greater precision than the 1% steps typically used because I have a set of sequences that are very similar to eachother. Specifically, I need to cluster OTUs at a level of 99.5% similarity or 0.5% dissimilarity. I see that the cluster command has an option for precision, whereby if the hard parameter is set to false then the cutoff can be modified. However, I am unclear if this would be applicable to my situation, as I do not know whether modifying the cutoff would also modify how each OTU is formed.

Thank you for any advice on this, and thanks again for such a great bioinformatics tool.

JK Cole

cluster(column=…, name=…, precision=1000)