To start off I have mothur v1.33.3 running on Centos6.4.
I have been trying to interface PICRUSt with mothur and one of the main steps involved is making a shared file to then use with make.biom command. I have used the make.shared command many times previously in prior versions but since updating to mothur v1.33.3 I am running into a segmentation fault when I run the make.shared command.
I first thought this could possibly be due to the fact that I am using greengenes OTU IDs as this was the only thing that I had changed in my pipeline between versions of mothur to accommodate the need for them in PICRUSt (previously I had been using RDP to classify the sequences). I also went back and ran the same pipeline in mothur v1.32.1 and didn’t get a segmentation fault. I did a quick search through the forum and people in the past who ran into segmentation faults seem to have too little memory on their workstations but I have 128GB of RAM on the blade in which mothur is running so I feel I can confidently rule memory out as the issue besides the fact that the previous version of mothur was able to run the make.shared command just fine.
I would be glad to provide my group and list files for debugging reasons if necessary. Thanks in advance for the help.
I went back today and ran our basic pipeline (using RDP to classify the sequences) on the new version of mothur (v1.33.3) and everything ran fine, even the make.shared command. This really does lead me to believe that the issue somehow lies where I am classifying the sequences the green genes OTU database versus the RDP database I have regularly used in my pipeline in the past.
I changed the file names of the list and group files in my unix environment for the make.shared command and the segmentation fault stopped occurring. As silly as this sounds this actually makes sense because the only difference in the two pipelines I had mentioned above were the naming of the files because when I classified the sequences with greengenes versus RDP it didn’t change the resulting fasta, group, and names files. All that it did differently was output a different taxonomy file. I also double checked to make sure that no contaminants were actually removed in to prohibit any differences arising from that step in my pipeline. The only real difference was the naming of the files and for some reason the name I gave the files which were classified with greengenes didn’t fly mothur.
I think I am going through the same kind of errors with chimera.uchime and classify.seqs.
I had segmentation faults chimera.uchime half of the times i would try to run it last week, and while classify.seqs was working well some days ago, i get segmentation faults today. (I am running mothur v.1.33.3)
I could not find out what was the problem but im gonna check about the file names.
If you could update to the most recent version of mothur (1.34.4) and try again and then start a new thread on the forum that would be great.