Fig. 11From: Similarity-assisted variational autoencoder for nonlinear dimension reduction with application to single-cell RNA sequencing dataEmbeddings from retina dataset. VAE and CVAE are applied to the retina dataset. From a and c, we find that the retina dataset is subject to a covariate effect, which causes additional groupings beyond the cell types. As shown in b and d, after the correction with CVAE, the covariate effect is well resolved. In c and d, the labels 0 and 1 represent the batch IDBack to article page