Fusion of Multimodal Data: A Network Medicine–Based Framework for Next‑Generation Therapeutic Target Discovery
##1: Bottlenecks in Drug Development and Limitations of Single‑Data Analyses Traditional therapeutic target discovery has relied on single‑omics data or specific biological features. However, complex diseases such as cancer exhibit high inter‑cellular heterogeneity and multilayered molecular networks, making it difficult to address high failure rates and enormous development costs with a one‑dimensional approach. In diseases with a complex microenvironment, such as clear cell renal cell carcinoma (ccRCC), a precise system capable of integrative interpretation of multimodal omics data has been urgently needed.
##2: Network Medicine Machine Learning: Synergy of Four Core Data Types The research team built a machine‑learning framework that fuses single‑cell transcriptomics (scRNA‑seq), bulk multi‑omics, genome‑wide CRISPR screening, and protein‑protein interaction (PPI) networks into a single model. This approach mutually compensates for the noise inherent in each dataset and prioritises disease‑specific targets from a systems‑level perspective. Rather than relying solely on gene expression levels, the framework evaluates the strategic value of each gene within the biological interaction network using machine learning.
##3: ENO2 and LRRK2: New Target Discovery and Drug Repurposing Success Applying the framework to renal cell carcinoma reproduced known targets and identified five novel candidates. Experimental validation showed that ENO2 inhibition produced the strongest anti‑cancer effect, and notably, the LRRK2 inhibitor currently in phase III trials for Parkinson’s disease emerged as a repurposing candidate for renal cancer. This represents a tangible achievement that can dramatically reduce the time and cost associated with de‑novo drug development.
##4: Generalisable Scalability and the Future of Data‑Driven Drug Discovery The proposed approach is not limited to a single disease; it offers immediate scalability to a wide range of refractory conditions. By moving beyond single‑feature heuristics and integrating multimodal data through machine learning, this strategy is expected to increase success rates and build cost‑effective pipelines for drug development. Ultimately, it will accelerate the era of precision medicine in which data dictate the direction of therapeutic innovation.
The high cost and attrition rate of drug development underscore the need for more effective strategies for therapeutic target discovery. Here, we present a network medicine-based machine learning framework that integrates single-cell transcriptomics, bulk multi-omic profiles, genome-wide CRISPR perturbation screens, and protein-protein interaction networks to systematically prioritise disease-specific targets. Applied to clear cell renal cell carcinoma, the framework successfully recovered established targets and predicted five therapeutic candidates, with subsequent in vitro validation demonstrating that among these, ENO2 inhibition had the strongest anti-tumour effect, followed by LRRK2, a repurposing candidate with phase III Parkinson's disease inhibitors. The proposed approach advances target discovery by moving beyond single-feature, single-modality heuristics to a scalable, machine learning-driven strategy that is generalisable across diseases.
This dataset exemplifies a classic hybrid study that combines wet‑lab CRISPR screening with dry‑lab machine‑learning network analysis. It provides precise guidelines on how to weight and integrate heterogeneous datasets when developing AI‑driven target discovery models.