#1 · Systems biology

Network models nominate interventions from omics and screening data

The shared thread is network-structured modeling that converts omics or perturbation data into ranked candidates for intervention. A differential microbial gene-family network framework with pathway enrichment is applied to IBD, diabetes and atherosclerosis stratified by sex. A dynamical model over gene deregulation networks, linking N, T and NT markers, is proposed to predict bidirectional reversal of cancer phenotypes. RepurNet screens compounds against reverse disease gene signatures and builds protein-interaction disease modules to prioritize repurposing candidates, while iDMET+ uses differential metabolomic profiles as metabolite sets. What is new relative to descriptive network characterization is the emphasis on nominating interventions. Within this set, validation is reported only for iDMET+, on renal, colorectal and small cell lung cancer data; the preprints describe predictions without reported validation.