Entering edit mode
12 months ago
ha2606
•
0
I am trying to assemble a metatranscriptome sample (~83M reads) using rnaSPAdes (v 3.13.0), and I am running out of memory on these runs.
After running the following code:
rnaspades.py -m 1500 -t 8 -o rnaSPAdes_output_37 --pe1-1 SM037_R1.fastq --pe1-2 SM037_R2.fastq
I get the processor error:
slurmstepd: error: Detected 1 oom-kill event(s) in StepId=19919309.batch cgroup. Some of your processes may have been killed by the cgroup out-of-memory handler
And the rnaSPAdes error:
== Error == finished abnormally, err code: -9
I have used these parameters to successfully assemble a metaT sample of similar size (~75M reads), and the memory limits set here should be adequate. I am worried there may be an issue with my input file-- it was pre-processed using standard techniques with Trimmomatic and FastQC, and I have been unable to find anything unusual about it.
Does anyone have experience with an error of this nature?
Thank you in advance!
Every dataset is different and it is possible that you need more RAM with this particular one. What is
-m 1500is that 15 G or 1.5G? Allocate more RAM and keep trying until theoom-killerror goes away and the process works (or you get some other error).Thanks for your response! The memory limit in spades is set in Gb, so it should be 1500 Gb.
Since you appear to be using SLURM are you asking for an identical amount in your
sbatchcommand or SLURM script? Probably not and that is what is likely causing the process to get killed. You will need to use--mem=NNgin your SLURM directive (along with appropriate core request).Yes, I was asking for a lower amount of RAM in my slurm script (1000 Gb) than in the spades code. I had assumed the assembly would not hit this limit, but it's definitely possible that this was the problem. Thank you for your assistance!