HUMAnN - RuntimeError: can't allocate lock

Microbial community profiling HUMAnN
1

Hello,

HUMAnN v3.6

I was running HUMAnN on some samples and noticed that it threw error msgs as below:

g__Streptococcus.s__Streptococcus_cristatus: 4215 hits
g__Streptococcus.s__Streptococcus_constellatus: 1974 hits
g__Lactobacillus.s__Lactobacillus_zeae: 35 hits

Total gene families from nucleotide alignment: 132161

Unaligned reads after nucleotide alignment: 43.7616719737 %


Running diamond ........


Aligning to reference database: uniref90_201901b_full.dmnd

Total bugs after translated alignment: 60
Exception ignored in thread started by: <bound method Thread._bootstrap of <Worker(Thread-1, initial daemon 23448461444864)>>
Traceback (most recent call last):
  File "/cm/shared/apps/HUMAnN/3.0/lib/python3.9/threading.py", line 930, in _bootstrap
  File "/cm/shared/apps/HUMAnN/3.0/lib/python3.9/threading.py", line 959, in _bootstrap_inner
  File "/cm/shared/apps/HUMAnN/3.0/lib/python3.9/threading.py", line 948, in _set_tstate_lock
RuntimeError: can't allocate lock

I have run humann_test and it went well. The sample has about 130 million reads.

Here is my HUMAnN command

humann --input my_sample_fastq \
        --threads 12 \
        --memory-use maximum \
        -o sample_name_humann_out \
        --metaphlan-options "--bowtie2db <local_bowtie2_db_path> --offline"

I’m just not sure how to address the error. Could you please help me with the error?

Thanks!
Joowook

2

This is a system-level error related to not being able to create a new thread or manage memory when multithreading. Maybe 12 threads is too much for your system? I would try with a smaller number and see if the error goes away.

3

Hi franzosa,

Thanks for the input. I actually assigned 14 cores to a node and used 12 threads to run HUMAnN, but I still encountered the same error. Interestingly, when I downsampled the FASTQ files to 80 million reads, the job completed without any issues. I’m not sure how to get around the error yet, but for now, I plan to downsample any FASTQ files with more than 100 million reads.

Thanks!
Joowook