Fig. 10From: Deep clustering of small molecules at large-scale via variational autoencoder embedding and K-meansThe proportion of variance explained based on the number of PCs. The first 50 PCs explained 100% of the variance of the dataBack to article page