Fig. 1From: N-of-one differential gene expression without control samples using a deep generative modelThe bulk deep generative decoder. The latent space is parameterized with a Gaussian Mixture Model (left). A generative neural network trained on the GTEx database maps the latent space to the data space. Altogether, the model learns the sample representations, the mixture model, and the gene expression distribution across bulk tissues, as illustrated in the heatmap (right)Back to article page