MIASA-prediction for scaled distances.

A: Families of distributions with max-scaled histograms representations, UMAP projections of predicted (in separate panels 1-4) versus convex hulls of the data points belonging to the different distributions: Poisson (Poi₁ to Poi₅), Normal (N₁ to N₅), Pareto (Pa₁ to Pa₅), and Uniform (U₁ to U₅). B: Correlated variables with (1/2)-scaled association distance, UMAP projections of predicted (in separate panels 1-10) versus samples of bivariate normal distributions, first dimensions (N_1X to N_10X) and second dimensions (N_1Y to N_10Y). C: Two-Gene Regulation Network with (1/2)-scaled similarity in gene A, UMAP projections of predicted (in separate panels 1-4) versus true gene regulation patterns between gene A and gene B (convex hulls of data point representations): No-I A, No-I B, Mono-I A & B, and Bi-I A & B. True and False predictions are only evaluated for the pairs of genes belonging to the same models.

(TIF)