error while running fragment-insertion

[salma_M (Saliya)]

Hello,

I have same problem; I had done PICRUST2 analysis before on a group of my data using qiime2-2022.2 version few months a go using virtual box and everything went very well. Now it seems that qiime does not support running analysis using virtual box, so I installed Ubuntu and then Qiime2 2023.7 and picrust2 on my new laptop with 16 GB of RAM, however checking on the terminal show free space around 7GB ...
"(qiime2-2023.7) wsl@Salma:~/picrust2-combined-cecum-feces$ free
total used free shared buff/cache available
Mem: 7975196 560708 139064 3196 7275424 7111140
Swap: 2097152 0 2097152"
Unfortunately I am stuck at the Fragment insertion step when trying to run this command;
qiime fragment-insertion sepp --i-representative-sequences Combined-Cecum-Feces-rep-seqs.qza --i-reference-database picrust2_default_sepp_ref.qza --output-dir sepp_output --p-threads 1 --p-debug --verbose
Plugin error from fragment-insertion:

Command '['run-sepp.sh', '/tmp/qiime2/wsl/data/935d1ae2-7ea5-4f7c-bcf8-9b8d5eaba6fa/data/dna-sequences.fasta', 'q2-fragment-insertion', '-x', '1', '-A', '1000', '-P', '5000', '-a', '/tmp/qiime2/wsl/data/2028ee20-c65b-470e-9cf5-80e2b4e28d0d/data/aligned-dna-sequences.fasta', '-t', '/tmp/qiime2/wsl/data/2028ee20-c65b-470e-9cf5-80e2b4e28d0d/data/tree.nwk', '-r', '/tmp/qiime2/wsl/data/2028ee20-c65b-470e-9cf5-80e2b4e28d0d/data/raxml-info.txt', '-b', '1']' returned non-zero exit status 1.

See above for debug info. (The debug info is very long but I can save it and send it to you if it may help)
, and I keep getting this error Plugin error from fragment-insertion after 30-40 minutes or more:

So this is my question, could this error be related to a bug for the most updated qiime2 version? I tried to downgrade the install qiime but again I could not run this step, and got an error on rep.seqs.qza showing that it is generated on qiime2. 2023.7 version and it was not compatible for this picrust2 fragment insertion step. Please keep me posted. Thank you.

[Oddant1 (Anthony Simard)]

Hello @salma_M. If your laptop is meant to have 16gigs, this could mean that it has 2 8gig sticks and one isn't being recognized for some reason which would explain why your free command is only showing 8gigs of RAM. If you open up your system info screen, how much RAM does it say you have? Additionally, did you dual boot Ubuntu alongside Windows or another OS? If so, you can check how much RAM that OS shows.

If you run your QIIME 2 command then in another terminal. If you run /usr/bin/time -v {your_command} the output will show, among other things, the maximum amount of memory your command used. It seems likely you are running out.

[salma_M (Saliya)]

Hello @Oddant1 ;

Thank you for the reply. The system shows 16 GB of installed RAM with 15.7 GB of usable RAM. However, for an unknown reason, the system information shows only 8 GB of available physical memory. I assume that no more than half of the total system RAM would be available for such analysis. I have not dual-booted Ubuntu alongside Windows, but I am using it on WSL. I'm not sure if WSL behaves the same way or not.

Initially, I thought there might be a problem with the fragment insertion step in the latest version of Qiime2. But, as you mentioned, it is likely a RAM issue since my current analysis involves double the number of samples compared to the ones I did before. I may need to acquire external RAM. Thank you for your help.

Salma

[Oddant1 (Anthony Simard)]

Yeah that makes sense WSL only has access to half your RAM. It does sound like you will need to run your analysis on a system with more RAM.