The Compositional Perturbation Autoencoder (CPA) is a deep generative framework to learn effects of perturbations at the single-cell level. CPA performs OOD predictions of unseen combinations of drugs, learns interpretable embeddings, estimates dose-response curves, and provides uncertainty estimates.
Hello,
I have installed the CPA software but I found the results very different from the published.
I think it is due to that I have installed the latest version of scvi-tools.
Could you please provide the specific version of scvi-tools used in this software ?
Hello, I have installed the CPA software but I found the results very different from the published. I think it is due to that I have installed the latest version of scvi-tools. Could you please provide the specific version of scvi-tools used in this software ?
Best