How to incorporate replicates into the analysis in mothur?

The requirement to include replicates of samples when conducting experiments is rarely fulfilled in studies of microbial ecology (particularly those which utilise high-throughput sequencing techniques), although it is a basic requirement in most scientific disciplines including most fields in biology, and despite having been stressed out in the literature (e.g. Prosser 2010, EMI).

Nevertheless, with the use of tagged sequencing technique one can easily analyse replicated samples in parallel, yet this issue seems to be absent in the mothur tutorials.

I was wondering if people here have suggestions on how to incorporate the replicates in the different stages of sequence analysis in order to achieve greater statistical robustness.
Particularly I’m looking for suggestions that go beyond the trivial steps of merging the results or of manually examining whether the replicates have the same taxonomical patterns or diversity indices.
Rather I’m looking for ways to include replicates more explicitly, for instance when computing diversity metrics or testing hypotheses.

Suggestions are welcomed!


Well… the example data we provide in the SOP is part of a much much larger study where there were technical and biological replicates performed. One simple thing you could do with the output data you get from the SOP are t-tests/whatever to compare the diversity early vs. late in the 10 samples. I hesitate to go down that road because at some point it would turn into a Stats 101 course or a course on R.

thanks and sorry for the late reply.
No, of course I wasn’t expecting any stats or R introduction…
One thing for instance is that it took me a while to figure out that one simply incoroporates replication in the .design file by giving two samples the same name.
I think it’s worth mentioning in the SOP or somewhere else…