Issues about the installation of dependence about phamb HOT 5 CLOSED

And I check now with python=3.7 on Windows10, it does also seem to work:

conda create -c defaults -c conda-forge -c bioconda -n phamb snakemake pygraphviz python=3.7 cython scikit-learn=0.21.3

I added explicitly the list of anaconda channels in search order: -c defaults -c conda-forge -c bioconda

from phamb.

@joacjo @enryH Thanks for your kindly help!
Yeah, the installation of python=3.6 can work!
conda create -n phamb python=3.6 cython scikit-learn==0.21.3 snakemake pygraphviz
It can work, while "python=3.7" can not work in my CentOS 7.6 system.

Thanks again!

By the way, to my understanding, is this phamb workflow used to identify viral MAGs after binning by Vamb to obtain miscellaneous MAGs?

from phamb.

Hi Jiulong

Thank you for your interest in the method. I did some experimentation and experienced the same conflicts in the conda installation as you present here. Seems like the required scikit version is not compatible with Python v. 3.8.

Try to give this one a go, python 3.6 should not result in the same conflicts with scikit-learn and worked stable before.
conda create -n phamb python=3.6 cython scikit-learn==0.21.3 snakemake pygraphviz

Let me know if this works out.

Best,
Joachim

from phamb.

As python 3.6 is out of live, I guess it will be good to see if this can be updated..

from phamb.

@joacjo @enryH Thanks for your kindly help! Yeah, the installation of python=3.6 can work! conda create -n phamb python=3.6 cython scikit-learn==0.21.3 snakemake pygraphviz It can work, while "python=3.7" can not work in my CentOS 7.6 system.

Thanks again!

By the way, to my understanding, is this phamb workflow used to identify viral MAGs after binning by Vamb to obtain miscellaneous MAGs?

Great it worked for you Jiulong! I will update the dependency for python 3.6, for now.
As Henry points out, python 3.6 is getting outdated, so we will have to prepare the method for a more recent version of Python.

To your question, Yes phamb workflow requires Vamb clusters.tsv and is optimised for that binner.
So you will need to run Vamb on a metagenomic dataset in order to employ this method.

Wish you all the best with your research!

from phamb.