I know that scaled data must be used for PCA for example as it is based on variance maximization.
However I'm wondering if it's the same case for UMAP ? If the data are single-cell RNA seq, after normalization can we do UMAP and Louvain clustering ?
I tried both and get similar results, however the Louvain clustering seems to be more adequate on normalized data than on scaled data.
By adequate I mean the clusters are the same but some are split into two, which makes sens looking at other results (like transcription factor analysis).