Getting error in the GTDB step.

[ankurnaqib]

Hi,

I am getting an error while running the GTDB step. I am attaching the log file. It seems like Prodigal failed due to an error "AttributeError: module 'numpy' has no attribute 'bool'". Please help. Thank you.

gtdbtk-run.log

[bnovak32]

Which version of the workflow are you running? Can you also provide information on the environment in which you ran the command? Was it a docker/singularity container (if so, which image/version)? Or if you used a conda environment, please provide the result of the conda export -n environment_name command. I suspect that your environment has a mismatch in the versions of some of the python packages used for that step.

[ankurnaqib]

Hi. I am sorry to be so delayed on this. I received this error again. I am attaching gtdbtk-run.log file from another dataset. I am getting the same error. I am using the SW_MGIllumina_2.0.4 workflow. I am using conda to run this.

gtdbtk-run.log

Basically 99.6% of the steps are done and it is just stuck on the last 8 steps. The screen shows this:-

I am also attaching the snakemake file and the gtdb-tk.yaml file with this message.

gtdb-tk.yaml

Snakefile.zip

Any help would be really appreciated. Thank you.

[bnovak32]

My GitHub notifications are clearly lacking because I also missed your response earlier.

I think the issue is that the gtdb-tk environment file doesn't fully specify the versions of the dependencies, which has caused incompatibilities in the installed software. Many of the errors seem to be in references to numpy.bool, which was deprecated in numpy version 1.20.0 and removed in 1.24. The conda env we used in production for that version of the workflow has numpy 1.21.2 installed. Please try adding a line to your gtdb-tk.yaml file to include numpy=1.21.2 as a dependency. I am including the environment export for our gtdb-tk environment for reference:

gtdbtk_1.5.0_env.yaml