jtamames / SqueezeMeta

A complete pipeline for metagenomic analysis
GNU General Public License v3.0
346 stars 81 forks source link

Stopping in STEP1 -> 01.run_all_assemblies.pl. Program finished abnormally #846

Closed Shivabrahmam closed 1 month ago

Shivabrahmam commented 1 month ago

Died at /home/miniconda3/envs/SqueezeMeta/bin/SqueezeMeta.pl line 941.

I was performing the coassembly then it gave this error, earlier to this i have even performed the sequential analysis it did not show any error.

fpusan commented 1 month ago

If things worked well using the sequential mode then it looks like you don't have enough RAM for the coassembly.

Shivabrahmam commented 1 month ago

I have around 100 GB of free RAM

fpusan commented 1 month ago

This may look like a lot but it is often far from enough depending on the number of samples, their depth and the complexity of your metagenomes. Can you share the syslog file inside the project directory with us?

Shivabrahmam commented 1 month ago

syslog.txt

fpusan commented 1 month ago

From the file I can see that MEGAHIT is failing with Exit code -9 which happens indeed due to a lack of memory (maybe you are working in a cluster and didn't ask for the right amount?) https://github.com/voutcn/megahit/issues/272#issuecomment-643017856

Shivabrahmam commented 1 month ago

Thank you for the help, I will try to run with the few samples and check for the results

Shivabrahmam commented 1 month ago

It worked with decreasing the number of samples as input, but I have one more doubt, in results I could only see the assembly against bacteria, I want to know that will it not identify any other like protozoa, fungi, viruses, etc. Because I am working on plant meta genomics it is required to identify them as well.

fpusan commented 1 month ago

It should identify viruses and eukaryotes too, if present, although the taxonomic resolution will be poorer, and eukaryotes may also fail to assemble and/or annotate. What's your percentage of mapped reads?

Shivabrahmam commented 1 month ago

For all the samples it is below 50%, highest I have seen is 35% and rest all the samples are in between 20 to 30%

jtamames commented 1 month ago

That is rather low. You could consider using sqm_reads.pl instead, to work with reads

Shivabrahmam commented 1 month ago

Thank you, I will try it for once.