AutoClass (IMAGE) University of North Carolina at Charlotte Caption AutoClass integrates a classifier to a regular autoencoder, as to fully reconstruct scRNA-Seq data. a AutoClass consists a regular autoencoder and a classifier branch from the bottleneck layer. The input raw expression data is compressed in the encoder, and reconstructed in the decoder, the classifier branch helps to retain signal in data compression. The output of the autoencoder is the desired imputed data (see Methods for details). b t-SNE plots of Dataset 1 without dropout, with dropout, with dropout imputed by a regular autoencoder and AutoClass. Credit UNC Charlotte Usage Restrictions Available for publication. License Licensed content Disclaimer: AAAS and EurekAlert! are not responsible for the accuracy of news releases posted to EurekAlert! by contributing institutions or for the use of any information through the EurekAlert system.