Running palantir on batch-corrected data using different integration methods

[miloAhn]

Hi,

I’m planning to run Palantir on batch-corrected data (WT and KO groups) and noticed that there are many integration methods available, such as scVI, Harmony, and CCA/RPCA. I’ve read this issue (#104 (comment)), which mentions that the X_scVI can be used as input to compute diffusion components.

I’m wondering if embeddings generated from other integration methods can also be used as input for computing diffusion components. For example, I integrated my data in Seurat using RPCA, converted the Seurat object to an AnnData object, and saved the RPCA embeddings in .obsm. Can I use these RPCA embeddings stored in .obsm to compute diffusion components in Palantir?

Thanks for your help!