PacificBiosciences / pb-human-wgs-workflow-snakemake

Workflow for the comprehensive detection and prioritization of variants in human genomes with PacBio HiFi reads
BSD 3-Clause Clear License
38 stars 20 forks source link

snakemake version incompatible with --max-threads=256 #126

Closed lauragails closed 1 year ago

lauragails commented 1 year ago

Hello - I followed the tutorial here and am now set to run the pipeline. Specifically related to this issue, I ran the following command to make the conda environment:

mamba create -c conda-forge -c bioconda -n pb-human-wgs-111622 snakemake=6.15.3 lockfile=0.12.2 pysam=0.16.0.1 python=3

This worked fine!

But when running $DIR_EXECUTABLE/workflow/process_smrtcells.lsf.sh I get the error: snakemake: error: unrecognized arguments: --max-threads=256

This seems to be due to the snakemake version. Later versions of snakemake have this argument, as snakemake -h in the created environment doesn't have that argument. The latest version of snakemake does, though link.

I'm installing a later version now, but thought I'd also fill out a ticket (especially since there might also be issues downstream).

Thank you (and thank you for your help at ASHG)!

lauragails commented 1 year ago

(it looks like this flag is present in snakemake 6.15.5 - trying that now!)

edit: no matter what version I type in, snakemake installs 5.26.1 (checked with --version. It seems like this is in fact the problem, since there isn't even a manual for that version!

williamrowell commented 1 year ago

Is it possible that there's a snakemake outside of the conda env and there's a path issue? What does which -a snakemake look like?

lauragails commented 1 year ago

which -a snakemake /sc/arion/work/sloofl01/pb-human-wgs/envs/pb-human-wgs-debug/bin/snakemake /hpc/packages/minerva-centos7/py_packages/3.7/bin/snakemake

Good point about the outside - that version in py_packages is the old version that doesn't have a manual.

What I think is happening is that I have py_packages and python loaded on our hpc when I'm creating the environment, and something is getting confused. I'll try to remake the environment again without loading our hpc modules, and maybe it will work.

Likewise, let me know if you have other suggestions!

On Wed, Nov 16, 2022 at 4:41 PM Billy Rowell @.***> wrote:

Is it possible that there's a snakemake outside of the conda env and there's a path issue? What does which -a snakemake look like?

— Reply to this email directly, view it on GitHub https://github.com/PacificBiosciences/pb-human-wgs-workflow-snakemake/issues/126#issuecomment-1317704174, or unsubscribe https://github.com/notifications/unsubscribe-auth/ABVNCLIH32GTIPCGB72PRULWIVIHPANCNFSM6AAAAAASCU55ZM . You are receiving this because you authored the thread.Message ID: <PacificBiosciences/pb-human-wgs-workflow-snakemake/issues/126/1317704174@ github.com>

williamrowell commented 1 year ago

It's weird because your env snakemake is higher priority in the path than the module snakemake. I've seen this sort of thing before, though. Somewhere, the path is getting scrambled. Maybe the module is getting loaded a second time and ending up higher in the priority. If you can avoid loading hpc modules, that's definitely the way to go about it.

lauragails commented 1 year ago

it's working!! module unloading python and py_packages did the trick.

So far, all of the files are entering the dictionary etc. and relevant packages are being downloaded, jobs are being created etc. Will let you know if other issues come up.

Thank you again!

williamrowell commented 1 year ago

No problem, glad we found the solution.