wheaton5 / souporcell

Clustering scRNAseq by genotypes
MIT License
168 stars 46 forks source link

Dependencies and Conda Environment Difficulties #229

Closed ggruenhagen3 closed 7 months ago

ggruenhagen3 commented 7 months ago

First off, amazing package! I've been using it for years and I think it's brilliant. Onto my issue, I struggled for awhile to getting all the correct package versions in order to run souporcell. I used the conda environment supplied in this github, but when I ran souporcell I would get this TypeError from tensorflow. I was unsure what the error meant, but I figured it was a version issue. I think I found that when installing tensorflow using pip there was an error with absl-py telling me that I needed to upgrade python. Regardless, after some trial and error, I found that upgrading python to 3.7 then installing some of the packages listed in the conda environment separately, such as pip install tensorflow==1.14.0, mamba install tensorflow==1.14.0 and mamba install pyvcf==0.6.8, worked for me (additionally I faced some GLIBC kind of issues that I solved by modifying my $LD_LIBRARY_PATH, but I'm not sure if that's necessary after the other steps). I'm not sure if I'm the only one facing this issue, in which case, maybe there's no need to change the conda yaml file, but if there's anyone facing the same issue, I hope this helps!

System Specs: CentOS Linux 7 with x86-64 architecture, using supplied conda environment in the souporcell github (though I can't remember if there were any errors/warnings while creating the environment)