Fig. 2
From: N-of-one differential gene expression without control samples using a deep generative model
Representations for normal tissue. A PCA plot showing the clustering of 31 tissues in the latent space reduced from 50 dimensions to 2. Each dot corresponds to a training sample. B Matrix of associations showing the percentage of the samples assigned to a Gaussian component (x-axis) represented by each tissue (y-axis). The tissue types and component numbers are sorted for an optimal diagonal view. The numbers below the x-axis give the number of samples assigned to each component. The number of samples in each tissue is shown on the right side of the matrix