mothur

Tree.shared not running on multiple processors

I have some 16S MiSeq data that I am trying to run through the same pipeline as the MiSeq SOP. I have done this successfully until I get to the portion creating a tree. I’ve managed to use a high memory computer on the university cluster to overcome that hurdle, but Clearcut times out – potentially because it cannot run in parallel.

I recently started to try and use the ‘tree.shared’ function because it allows use of multiple CPUs. However despite having multiple processors available on the HPC cluster, I have not been able to make use of them using either the autodetect default or through explicitly specifying them with the ‘processors=’ argument.

Has anyone run into this issue? What types of workarounds have you used?

Thank you! This forum is wonderful.
-Chris

Also I should note that I am on v1.42.3, though I do now have 1.43.0 available but have not re-run the commands yet.

Hi,

tree.shared is really meant for building trees where the branches represent samples, not sequences. If you are running up against RAM limits with clearcut, I’d gently recommend skipping phylogenetic approaches. We have yet to see a case where a phylogenetic approach has disagreed with the analogous OTU-based analysis.

Pat

This topic was automatically closed 10 days after the last reply. New replies are no longer allowed.