Multi-Omics Integration of Transcriptomics and Metabolomics with Machine Learning Uncovers Novel Risk Factors for...

Area of Science:

• Neuroscience
• Genomics
• Metabolomics

Background:

• Alzheimer's disease (AD) is a complex neurodegenerative disorder with unclear pathogenesis, involving amyloid plaques, tau tangles, neuroinflammation, and synaptic dysfunction.
• Genetic, environmental, and lifestyle factors contribute to AD risk, but their interactions remain poorly understood.
• Advances in transcriptomics and metabolomics reveal gene expression alterations and metabolic disruptions in AD progression, necessitating integrated analytical approaches for biomarker discovery.

Purpose of the Study:

• To integrate genetically imputed whole blood transcriptomics and plasma metabolomics to predict cognitive performance (PACC3).
• To identify risk genes and metabolites contributing to cognitive performance prediction in AD.
• To characterize molecular signatures associated with cognitive performance in Alzheimer's disease.

Main Methods:

• A machine learning algorithm was employed to integrate transcriptomics and metabolomics data for cognitive performance prediction.
• The model was trained on the Wisconsin Registry for Alzheimer's Prevention (WRAP) cohort (N=1,046) and validated on the Wisconsin Alzheimer's Disease Research Center (ADRC) cohort (N=85).
• Feature importance analysis was conducted to identify predictive genes and metabolites associated with AD risk and resilience.

Main Results:

• The machine learning model demonstrated predictive performance in both WRAP (R²=0.311) and ADRC (R²=0.061) cohorts.
• Transcriptomic biomarkers like RIPK1, IL6ST, and BIN1 were associated with poorer cognitive performance, while UGP2, NDUFB5, and TMOD2 were linked to better performance.
• Predictive metabolites, including benzoate and imidazolelactate, mapped to AD vulnerability, while certain acyl-carnitine species aligned with metabolic resilience.

Conclusions:

• Integrated transcriptomics and metabolomics analysis shows promise for identifying AD biomarkers.
• Genes and metabolites related to inflammation, mitochondrial dysfunction, and lipid metabolism were key contributors.
• These findings support the value of multi-omics integration for characterizing AD and prioritizing biomarkers for future research.