This command crashes either my computer or just mothur. I’ve run it when having 100,000s of sequences, which crashes the computer and mothur crashes if I try to run only 47000 sequences. I’ve run the larger one on 2 different computers the lesser of them being a Intel Xeon X5650 with 48 GB or RAM, the other computer being better but I do not have the complete details.

I’m confused as why this is occurring.

How many OTUs are in the table you’re processing? The process behind the command is quite complicated and I’ve definitely noticed it slows right down once you start processing 1000’s of OTUs.

In the smaller one there is 90 OTUs.

How many samples are there?

The smaller one has only 2 samples, which I can understand cause issues but to crash the program?

The larger one has 5 samples.

I have another set of ~45 samples with a total of 3 million sequences that I haven’t run using this command concerned it one won’t work as I’ve tried to leave the one with more sequences over a full weekend with no success except an extremely sluggish computer. Or it crashes the computer if you attempt something else.

Is this just because it has either too few samples and my other ones have too many sequences?

I suspect the problem is that it’s trying to form more cluster than you have samples. This is an edge case that we hadn’t foreseen. I think this command really only makes sense if you have a decent number of samples (like 45). Can you try the 5 sample dataset with a max of 4 clusters (i.e. maxpartitions=4)?