We work with time series data and are interested in making a “global” set of otus for our system in order to cluster future data against our global otus and see how these otus behave over time/space. As far as I can tell, there is no way to use mothur to do this second clustering step, aka, clustering against a database. Is this correct? I am very aware this may hamper the detection of novel otus, but we feel that for our datasets (for a well-characterized environment) and our questions (how do particular OTUs change over time–e.g. over decades–in relation to other OTUs or in “absolute” abundance) could greatly benefit from using a database to cluster against. Do you know of a way in mothur to accomplish this?
You could make a reference taxonomy based on your reference OTUs and then use phylotype to compare your reads to the reference after running classify.seqs with something like method=blast or whatever you please.
I myself do not get why in this case it would jut not be better to re-cluster the entire read set over the whole timeseries. I feel that this is statistically more relevant and correct to have unbiased OTU clustering over time and follow up the progression.
What am I missing here?
FWIW, I agree with you FM.