How to use normalised expression


I would like to get some more information about the normalised expression of genes once the model has been trained.

Let’s suppose that anndata is my object over which I train my model.

  • Are the values in anndata.X the same as the original raw counts contained before training the model?

  • How does the function get_normalized_expression work? I have seen that it returns a dataframe containing the normalised expression of genes from the decoder, but how are these values calculated? Are they the same as the ones stored in anndata.X? Are they log-normalised?

  • I would like to plot the expression for few genes on the umap to see which clusters express my genes of interest. Would it be okay to plot the expressions of the genes obtained from model.get_normalised_expression(gene_list=my_genes) using default parameters? Or you would recommend using different parameters?

Thank you very much for your help!