Similarity Network Fusion: Multi-Omics Patient Subtyping

Multi-omics integration means combining genomic, transcriptomic, proteomic and other layers of measurement into one coherent picture of a patient or sample. This simulator implements Similarity Network Fusion (SNF): it builds one patient-similarity network per omics layer, then iteratively fuses them by letting each layer's network diffuse through the consensus of the others, so signal that is real in multiple layers strengthens while single-layer noise washes out. Toggle layers on and off, step through the fusion iterations, and watch a 3D force-directed graph of patients relax from an undifferentiated cloud into cleanly separated disease subtypes — exactly the effect SNF produced on real TCGA cancer cohorts by combining mRNA expression, miRNA expression and DNA-methylation similarity networks.