From Patient-Specific Networks to Actionable Drug Repurposing Strategies through Multi-Omics Data Integration
Year: 2026
Type: conference-paper
Abstract: The increasing availability of multi-omic data offers unprecedented opportunities to characterize disease mechanisms at molecular resolution. However, translating these heterogeneous datasets into clinically actionable knowledge remains a major challenge. Network medicine provides a systems-level framework for addressing this complexity by representing disease and therapeutic response as perturbations of interconnected molecular processes rather than isolated genes or proteins. In this talk, I will present two examples illustrating how network-based multi-omics integration can support patient stratification and therapeutic prioritization. First, using multi-omic profiles from 101 patients with lung adenocarcinoma, we developed an extracellular matrix (ECM)-guided stratification framework that identified four ECM grades with distinct clinical, mutational, and cellular characteristics. Patient-specific network models revealed enrichment of epithelial–mesenchymal transition and cancer stemness programs in higher ECM grades. Network-proximity analysis further predicted olaparib as an ECM-dependent therapeutic, whereas erlotinib showed an ECM-insensitive response; these predictions were subsequently validated experimentally in lung cancer models representing distinct ECM microenvironments. In a complementary study, we reconstructed 236 drug-modulated networks covering 70 drugs across five cancer cell lines by integrating multi-omic responses with drug targets and molecular interaction networks. Network-level comparisons revealed context-specific pathway perturbations, similarities between drugs not apparent from their annotated mechanisms of action, and evidence-supported candidate drug combinations. For example, I-BET-762 and lenalidomide emerged as the top-ranked combination in A375 melanoma cells, with topologically separated network effects and complementary pathway perturbations, consistent with prior evidence supporting combined BET inhibition and lenalidomide treatment. Together, these studies show how patient-specific and drug-specific network models can link molecular heterogeneity to therapeutic response and help prioritize context-dependent strategies for drug repurposing.
Author: Nurcan Tuncbag
Institution: Koç University
Language: English
Subfield: Molecular Biology
Domain: Life Sciences
Open Access status: gold
OpenAlex ID: W7204227355