phonograph.cluster() function doesn't seem to work correctly on Windows Subsystem for Linux

[danli349]

Hi
I run this PhenoGraph tutorial "tutorial_pbmc3k.ipynb"
But the “phonograph.cluster()” function can’t successfully calculate the communities, all the communities are zeros. But there is no error or warning.
Is there something wrong with this function?
IPython 7.20.0
jupyter_client 6.1.11
jupyter_core 4.7.1
jupyterlab 3.0.7
notebook 6.2.0
scanpy==1.6.1 anndata==0.7.5 umap==0.4.6 numpy==1.19.2 scipy==1.4.1 pandas==1.2.1 scikit-learn==0.24.1 statsmodels==0.12.1 python-igraph==0.8.3 louvain==0.7.0 leidenalg==0.8.1

Thanks
Dan

[hisplan]

We have tried the tutorial notebook again and it's working fine. Are you still having this issue?

[danli349]

I still can't run it through, the communities are all zeros:

[danli349]

[hisplan]

Strange. Are you using the dataset 3k PBMCs from a Healthy Donor (from 10x website)?

[danli349]

I am using the data in the "examples"
https://github.com/dpeerlab/PhenoGraph/tree/master/examples/data/filtered_gene_bc_matrices/hg19

[danli349]

It seems this problem only happens on Ubuntu (Windows Subsystem for Linux), I run the same tutorial "tutorial_pbmc3k.ipynb" on Windows 10, it can run through successfully.

[hisplan]

Sorry, we do not currently have a way to test this under Windows Subsystem for Linux. Did you check adata.X or adata.obsm["X_pca"] if these objects have non-zero values for example??

I'm also curious whether a much simpler example works on your environment or not:
https://github.com/dpeerlab/PhenoGraph/blob/master/examples/example-using-random-data-points.ipynb

[danli349]

Yes, this tutorial
https://github.com/dpeerlab/PhenoGraph/blob/master/examples/example-using-random-data-points.ipynb
can run through successfully on both Windows 10 and Ubuntu (WSL).
Thanks

[hisplan]

Okay. That's good to know. Just trying to isolate the problem here.

Do you think you can restart the kernel, rerun the notebook from scratch, save the adata right before the section 1.4 and send us both the notebook and adata file (.h5ad)?

[danli349]

For tutorial "tutorial_pbmc3k.ipynb" on Ubuntu(WSL):

[hisplan]

@danli349 so I loaded up your adata and just performed the clustering on different platforms:

• macOS Mojave 10.14.6
• Ubuntu 20.04
• CentOS 7.6

Yet, I couldn't reproduce the issue you're having. I think this is very specific to WSL. Aforementioned, we do not have that environment so cannot do any more testing at the moment.

Is there any particular reason that you want to use WSL for your analysis?

[danli349]

Great. Thanks a lot.