Network-phenotype key gene screening (NPKG) identifies therapeutically relevant targets for neurodegenerative diseases

Age-related macular degeneration (AMD) shares common pathogenic mechanisms with other neurodegenerative diseases, but conventional gene discovery often fails to capture the critical interplay between molecular networks and disease phenotypes. We here introduce Network-Phenotype Key Gene screening (NPKG; made accessible at http://www.genetictargets.com/NPKG ), a computational framework integrating gene co-expression topology with phenotype relationships. Its workflow constructs scale-free networks, identifies phenotype-correlated modules, and refines key genes using network metrics and regularised linear regression. Applied to AMD, the framework outperformed traditional methods and prioritised 11 high-confidence genes. This gene set was verified by GeneAgent (a language agent utilising a domain-specific knowledgebase) and systematically validated through literature, network, and single-cell analyses. These validations confirmed five known loci ( SERPING1 , C1S , CFB , CD44 , and B2M ) and three understudied hubs ( S100A11 , HLA-DRA , and FCER1G ) as key regulators of choroidal immune dysregulation. Crucially, it exhibited generalisability by prioritising neurodegeneration-associated genes ( NEBL , SCAND2 , FAM108A1 , and GAPDH ) in Alzheimer’s disease with higher precision than benchmarks. Our results establish NPKG as a scalable strategy for enhancing mechanistic insights and accelerating target discovery. The prioritisation of biologically coherent targets confers direct therapeutic relevance; the NPKG-derived gene set, including complement factors for protein replacement and immunomodulators for microenvironment reprogramming, is suited for further mRNA therapeutic development, advancing translational medicine for neurodegenerative diseases.