vsearch + otu.hierarchy= your file did not include the label ....

new to the forum so forgive me if this is not the propper way to query.

I have a large dataset (1 million unique seqs, 1k samples).

I need to use otu.hierarchy for my research (various thresholds),
however it was not possible to generate the distance file to cluster
(well, I wasn´t allowed actually by the computer cluster facility
people as it seemed it was going to take forever and generate a ca.
20T file). After reading your peerJ paper I turned to (previously
following the Miseq SOP) vsearch:

Your file does not include the label
0.03. I will use the next smallest distance.
Your file does not include the label
0.25. I will use the next smallest distance.

So it seems that my list file only contained clusters at 0.255.
I then re-run the same cluster command without cutoff, it said 'You did not set a cutoff, using 0.03.'and the apparently only produced OTUs at 0.03

So, is vsearch going to generate OTUs at a single threshold (hence the
cutoff has another meaning as in dist)? or else what am I missing?

If this was the case (I mean vsearch working at individual
thresholds), is there a way I could achieve performing the
otu.hierarchy approach at various labels (appart from running vsearch
at all levels desired and the writting a dedicated script to link OTUs
at different levels)? is it possible to tune otu.hierarchy to link
OTUs at several levels at the same time?

(and thanks for mothur too!)

Have a look at the subsection “Finer points / Missing Distances” in the cluster wiki. (http://www.mothur.org/wiki/Cluster)
I think that might be your answer.

The vsearch methods just cluster for a specific distance. You would need to run the clustering for the distances you want to input to otu.hierarchy.

mothur > cluster.split(…, cutoff=0.03)
mothur > cluster.split(…, cutoff=0.25)

manually copy OTU 0.25 clusters line into 0.03 file.

mothur > otu.hierarchy(list=name.unique.good.filter.unique.precluster.pick.pick.dgc.unique_list.list,