Need to handle NaN values in adata

[jianxinwang]

Hi, Thanks for developing this tool. I just give it a try and it erred out with this message when I run this command: "usage, scores = tcat.fit_transform(adata)". The message is: ValueError: Input X contains NaN. I think rather than solving the problem by the end user, it is better to be handled in the fit_transform code. Jason

[dylkot]

Hi, yes potentially starCAT could just set all the output usages and scores for the cell containing the NaN to NaN but return results for the rest of the cells. I'll look into this.

[jpark27]

Dear @dylkot

Hi, I have also encountered same issue with starCAT while running with in-house data as follows:
/project/simmons_hts/jpark/anaconda3/envs/starcat_env/bin/starcat --reference "TCAT.V1" --counts "/project/simmons_hts/jpark/0_tools/starCAT/IBD_CD4_multiome/Xenium/matrix.mtx.gz" --output-dir "/project/simmons_hts/jpark/0_tools/starCAT/IBD_CD4_multiome/Xenium/" --name "Xenium_default"

Any suggestion how to fix the issue?

best wishes,
J