search

lsteuernagel / mapscvi

Map single cell expression data into a reference scvi latent space and reference umap using R and Seurat. Use included objects to map new data to HypoMap.
Other
1 stars 3 forks source link

Can not import hypoMap_harmonized_scVI_model #8

Open LILI-0000-0002-8173-7367 opened 1 month ago

LILI-0000-0002-8173-7367 commented 1 month ago

Hello,

I would like to use the trained hypoMap_model to annotate my data following the example notebook link. However, I cann't load the model, using _scvi.model.SCVI.prepare_query_anndata(adata_query, modelpath).

The error information is "ModuleNotFoundError: No module named 'pandas.core.indexes.numeric'". As @martinkim0 mentioned at https://github.com/scverse/scvi-tools/issues/2387. The issue is because of the model was generated when Pandas <2.

I would like to know if there is any chance you have an update hypoMap_model file using new version of scVI? The current scvi-tools version is 1.2.0.

Thank you very much.

Best regards, Li Li

lsteuernagel commented 2 weeks ago

Hi,

Yes, I already ran into this issue as well, so far my only workaround is to use an older environment with pandas version < 2.

I have not re-trained a new model on the existing data yet, but will consider doing that and update you when I get to do it.

Reproducibility of old UMAPs could be a problem with a new model, but label transfer with scArches etc. should probably work just as well.

Best regards,

Lukas