Abstract:
Neurodegenerative diseases (NDDs) such as Alzheimer’s disease, Parkinson’s disease, and amyotrophic lateral sclerosis are among the most debilitating disorders worldwide, characterized by progressive neuronal loss, molecular heterogeneity, and limited therapeutic options. Despite extensive research, the early detection and mechanistic understanding of these diseases remain major clinical challenges.

This study employs a comprehensive multi-omics framework integrating transcriptomic, proteomic, and metabolomic datasets from human post-mortem brain tissues and cerebrospinal fluid. By leveraging artificial intelligence (AI) and machine learning algorithms, including random forest classifiers and deep autoencoders, data from over 2,000 patient samples (ADNI, GEO, and AMP-PD) were analyzed to identify robust biomarkers and molecular subtypes.

Results demonstrate 31 shared molecular networks dysregulated across NDDs, prominently involving mitochondrial impairment, autophagy dysfunction, and neuroinflammatory pathways. The AI-based diagnostic model achieved 92% classification accuracy for distinguishing early Alzheimer’s disease from age-matched controls. Key hub genes—LRRK2, TREM2, and SYNGR3—were identified as central regulatory nodes. In-silico drug repurposing further suggested metformin and rapamycin analogs as potential modulators of these targets.

In conclusion, this research underscores the potential of AI-driven multi-omics integration in unveiling cross-disease biomarkers and accelerating precision diagnostics in neurodegenerative disorders. Future work aims to validate these findings through clinical cohorts and digital neurophenotyping.

In conclusion, this research underscores the potential of AI-driven multi-omics integration in unveiling cross-disease biomarkers and accelerating precision diagnostics in neurodegenerative disorders. Future work aims to validate these findings through clinical cohorts and digital neurophenotyping.