Hello! I am looking for some insight into an issue I keep running into. I am at the step of pre.cluster and I am running this code in a computing cluster. I am maxing out the memory available at ~500G yet I am continuously experiencing a 7040 Segmentation fault (Exited with exit code 139). Please correct me if I am wrong, but I believe this has to do with the RAM capacity? For reference, following filter.seqs, my total # of unique seqs was 19887899. Also, I am not sure if this is normal based on the number of sequences, but most codes up to this point have taken 3-7 days to run.
Does anyone have any insight into memory requirements for this sequence size or any other thoughts?
Thanks so much!
Lindsay
Hi Lindsay - what region are you sequencing? You might want to check out this blog post…
Pat
Hi Pat, I have paired-end 2x300 MiSeq 16S V4 data. I am (now) aware of the major drawbacks of this chemistry and have been continuously struggling with the processing as a consequence. Based on your forum post, understand that I could take a phylotype-based approach for classification, but I am still in the ‘processing improved sequences’ phase and keep these roadblocks. Any advice would be much appreciated!
Lindsay
are you to pre.cluster yet? you could increase the diffs argument to equal you length divided by zero and round down to an integer value.
pat
Hello! Yes, pre.cluster is where the fault is occurring, but it takes about 7 days after the code first began to error out. Given my chemistry, most of the sequence lengths are ~400 so I will try diff = 4 and see if that helps. Thank you!
This topic was automatically closed 10 days after the last reply. New replies are no longer allowed.