(ERR): bowtie2-align exited with value 1

Question

[StacyG]

I know there are a ton of messages about this already but I've looked through them and can't find a solution. I'm running bowtie2 on a HPC. I have successfully built the index using

bowtie2-build mm39.fa mm39

And am now trying to align a .fastq file to it. My command is:

bowtie2 --very-sensitive -x reference/mm39 -1 fastqFiles/Read1.fastq -2 fastqFiles/Read2.fastq -S alignedFiles/output.sam

and I'm getting an error:

readU: Success
(ERR): bowtie2-align exited with value 1

My .fastq files are VERY large. The first few lines of Read1.fastq:

@A00564:637:HT3M3DSX5:1:1101:1090:1000 1:N:0:ACTCAGAC
NCTCTAGCCTCCCACTCCCACCCTTCCCCTCCGGGAGACCTTCCAAGAGGTGAAGCCACATCCTTAGCAAACTCTCCAGTGCCTGTCTCTTATACACATCTCCGAGCCCACGAGACACTCAGACATCGCGTATGCCGGCGTCGGCTTGAA
+
#FFFFF:FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF:FFF:FF:FF:FFFFF:FFFFFFFFFFFFFFFFF,FFFF:FFFFF:FFFFFFFFFFFFFFFFF:FFFFFFF::F:FFF,:F:FF,,,F,,F:,,F,:
@A00564:637:HT3M3DSX5:1:1101:2410:1000 1:N:0:ACTCAGAC
NTCCCAAGGTGTCACCCTTCTAGAGGAAGGGAAGGTGCGGGCACACCCAAACCTGTCTCTTATACACATCTCCGAGCCCACGAGACACTCAGACATCGCGTATGCCGTCTTCTGCTTGAAAAGGGGGGGGGGGGGGGGGGGGGGGGGGGG
+
#FFFFF:FFFFFFFFFFFFFFFFFFFFFFFFF:FFFFFFFFFFFFFFFFFFFFFF:FFFF:FFFFFFFFFFFFFFFFFFFFFFFFF:FFFFFFF,FF,:FF:FFFFFFF:F:FFF:FF:,:,,,:FFFFFFFFFFFFFFFFFFFFFFFFF
@A00564:637:HT3M3DSX5:1:1101:6732:1000 1:N:0:ACTCAGAC
NTCATTGTCTAGTGTTATGGGTTAAATGTAAAATGTCCCCAGAGTCTCCTGTGCTTGAGCTGGTGGTGCTATTCTAGGAGGCTGTGGAGGCTTAGGAGATGGGGGGGGGGTGGTCACTGGAGAGAACTTTGGGGTTTTAGATCTATCCAC
+

Like I said, I have googled this and read the previous post with the same error and can't find anything. Any help would be appreciate.

Thanks,

[GenoMax]

I'm running bowtie2 on a HPC

Are you using a job scheduler? Are you running the command directly on a node?

-1 fastqFiles/Read1.fastq -2 Read2.fastq 

Are your reads in two separate locations or is that simply an error with copy/paste? You seem to be missing fastqFiles/ before Read2 in the command.

[StacyG]

I'm using slurm, so not directly on a node.

I checked the fastqFiles is in my script...just a type above. I'll fix that.

[GenoMax]

Are you assigning enough memory to this job via SLURM? If you are using SLURM was the only error message one you poste above?

[StacyG]

I've choosen different queues that are available and they all return the same error. Yes, that's the only error...nothing about memory. And I get it almost instantly after submitting the job.

[Mensur Dlakic]

There are dozens of posts on this site with bowtie2-align exited with value 1 in the title or somewhere in the thread.

https://www.google.com/search?q=bowtie2-align+exited+with+value+1+site:www.biostars.org

If none of those apply to you, there is a good chance that no Biostars reader knows how to solve this problem.

Answer 1

[ATpoint]

Here are the steps to narrow it down, for example using an interactive slurm session on a node:

Please check the log of the indexing. Did it complete without error and which files were produced, check with ls in the index folder.

Use head -n 40 on a fastq file pair and make a tiny example dataset. Execute the align command manually on these files. See how it goes. Does it get killed instantly?

Check that relevant paths indeed exist and are accessible from the nodes tjat slurm submits to.

If both above steps are ok then run manually on a full dataset, check with top if it runs. Memory should not be an issue, it uses really little melory for genomes like this. Fastq size should also be no issue, it just runs longer.

Also submit a job with slurm ratjer than interactively on that tiny test dataset.

Please show what these steps resulted on so one can natrow it down.

[StacyG]

Thank you! I followed the steps above and think I found the problem. It was throwing the error from me just calling the bowtie2 command. I switched to a different environment on the HPC and it seems to be running now. At least it's running for about 3 minutes and hasn't thrown the error yet. I'll keep checking and let you know. Really appreciate the help.

[StacyG]

Now I got error: (ERR): bowtie2-align died with signal 11 (SEGV) (core dumped).

I think I need to reach out to the HPC people...this might be a problem on their end. I was assuming it was something I was doing wrong.

[GenoMax]

This signal is for a memory fault. You need to find out how to run SLURM jobs asking for a set amount of memory (--mem=NNg option). There is a default amount that is assigned to jobs based on local configuration and your job likely exceeded that.

[ATpoint]

Yes, seems like inappropriate resource allocation so slurm kills it. Indeed sit with a HPC person to go through your submission script once to learn what all needs to specifically be set for such a job. Sounds solved to me towards the underlying problem, the rest is now getting familiar with specific slurm settings.