10.1101/2023.05.22.541678

A variational autoencoder trained with priors from canonical pathways increases the interpretability of transcriptome data