Vsearch dereplicate seqs error

[Zach_Burcham]

Hello,

I am trying to analyze some pre-filtered FASTA seqs and I got my seqs imported using the Qiime1 demux method and made my seqs.qza file. But when I try to dereplicate using this:

qiime vsearch dereplicate-sequences --i-sequences seqs.qza --o-dereplicated-table table.qza --o-dereplicated-sequences rep-seqs.qza

I get the following error:

Plugin error from vsearch:

Command ‘[‘vsearch’, ‘–derep_fulllength’, ‘/tmp/qiime2-archive-7lqgsqxg/f63f6a25-26e2-47e9-98f3-a2bb86365fba/data/seqs.fna’, ‘–output’, ‘/tmp/q2-DNAFASTAFormat-fl9gk__5’, ‘–relabel_sha1’, ‘–relabel_keep’, ‘–uc’, ‘/tmp/tmp_3xisfcy’, ‘–qmask’, ‘none’, ‘–xsize’]’ returned non-zero exit status -9

Debug info has been saved to /tmp/qiime2-q2cli-err-z0xejn24.log

Can anyone tell what is causing this error???

EDIT: I actually found the issue. I was having an issue with the amount of RAM my VM was allowed to use. Now I just have to figure out how to create a MAFFT alignment with about 11M seqs!!

Thanks!
Zach

[colinbrislawn (Colin J Brislawn)]

Good work, Zach!

I'm glad you got this figured out so fast and shared your answer on the forums.

@Zach_Burcham

EDIT: I actually found the issue. I was having an issue with the amount of RAM my VM was allowed to use. Now I just have to figure out how to create a MAFFT alignment with about 11M seqs!!

MAFFT tips: how about clustering your reads at 97% or 99% or 100% (dereplicating) to remove very similar reads. This should reduce your number down from 11M!

[Zach_Burcham]

Thanks, @colinbrislawn ! I am noticing when I try to dereplicate my seqs I am getting a good number of unique seqs only found 1 time. I see in the vsearch there is a command to make the unique cutoff 2, but is there a way to use this in the Qiime2 plugin?

[colinbrislawn (Colin J Brislawn)]

Hello Zach,

Getting a lot of singletons is super common, both because of the evenness of many biological communities and also because of the error profile of most sequencers.

Based on the documentation for vsearch dereplicate-sequences, it does not look like qiime can pass a --minuniquesize. You could however, use vsearch directly to do dereplication or use the qiime command cluster-features-de-novo to group your features together.

Colin

[thermokarst]

Thanks @colinbrislawn! For posterity, here is a link to the open issue created about exposing the --minuniquesize parameter in q2-vsearch.