Fig. 1From: Compressing gene expression data using multiple latent space dimensionalities learns complementary biological representationsOverview of the BioBombe approach. We implemented BioBombe on three datasets using five different algorithms. We compressed input data into various latent dimensionalities. We calculated various metrics that describe different benefits and trade-offs of the algorithms. Lastly, we implemented a network projection approach to interpret the compressed latent features. We used MSigDB collections and xCell gene sets to interpret compressed featuresBack to article page