Fig. 4From: Compressing gene expression data using multiple latent space dimensionalities learns complementary biological representationsAssessing gene set coverage of specific gene set collections. Tracking results in TCGA data for three gene set collections representing transcription factor (TF) targets (C3TFT), Reactome pathways (C2CPREACTOME), and cancer modules (C4CM). a Tracking coverage in individual models, which represents the distribution of scores across five algorithm iterations. b Tracking coverage in ensemble models, which represents coverage after combining all five iterations into a single model. The size of the point represents relative enrichment strength. c Tracking coverage in all models combined within k dimensionalities. The number of algorithm-specific unique gene sets identified is shown as bar charts. Coverage for all models combined across all k dimensionalities is shown as a dotted navy blue lineBack to article page