Skip to main content
. Author manuscript; available in PMC: 2018 Jan 1.
Published in final edited form as: Pac Symp Biocomput. 2018;23:80–91.

Fig. 2. Samples encoded by a variational autoencoder retain biological signals.

Fig. 2

(A) t-distributed stochastic neighbor embedding (t-SNE) of TCGA pan-cancer tumors with Tybalt encoded features. (B) t-SNE of 0-1 normalized gene expression features. Tybalt retains similar signals as compared to uncompressed gene expression data. (C) Full Tybalt encoding features by TCGA pan-cancer sample heatmap. Given on the y axis are the patients sex and type of sample.