Integrative Multi-Omics and Multivariate Longitudinal Data Analysis for Dynamic Risk Estimation in Alzheimer's...

Area of Science:

• Neuroscience
• Biostatistics
• Genomics

Background:

• Alzheimer's disease (AD) presents heterogeneous cognitive and functional impairments.
• Accurate AD progression assessment requires integrating diverse data, including neuropsychological tests and multi-omics (metabolomics, lipidomics).
• Challenges exist in utilizing high-dimensional, heterogeneous omics data for dynamic dementia risk estimation.

Purpose of the Study:

• To develop a novel joint-modeling framework for integrating multi-omics and longitudinal data.
• To enable dynamic risk evaluation for Alzheimer's disease progression.
• To address challenges in omics data utilization for dementia risk.

Main Methods:

• Combined multi-omics factor analysis (MOFA) for dimension reduction and feature extraction.
• Employed a multivariate functional mixed model (MFMM) for longitudinal outcome modeling.
• Integrated MOFA and MFMM into a joint-modeling framework.

Main Results:

• The proposed integrative joint modeling framework effectively combines multi-omics and longitudinal data.
• Demonstrated the framework's efficacy through extensive simulation studies.
• Successfully applied the model to the Alzheimer's Disease Neuroimaging Initiative (ADNI) dataset.

Conclusions:

• The novel joint-modeling framework facilitates dynamic evaluation of dementia risk.
• This approach leverages both omics and longitudinal data for improved AD progression assessment.
• The method shows practical utility in real-world datasets like ADNI.