Omics‐Imaging Data Integration via Mediation Analysis with High‐Dimensional Exposures and Mediators
Yi Zhao, Lexin Li- Psychiatry and Mental health
- Cellular and Molecular Neuroscience
- Geriatrics and Gerontology
- Neurology (clinical)
- Developmental Neuroscience
- Health Policy
- Epidemiology
Abstract
Background
Multimodal technologies have transformed AD research in recent years, by collecting different types of data from the same group of subjects and enabling the investigation of complex interrelated mechanisms underlying AD development.
Method
In this article, a mediation analysis approach with high‐dimensional exposures and high‐dimensional mediators is proposed to integrate data collected from multiple platforms. The proposed method combines principal component analysis with penalized least squares estimation for a set of linear structural equation models. The former reduces the dimensionality and produces uncorrelated linear combinations of the exposure variables, whereas the latter achieves simultaneous path selection and effect estimation while allowing the mediators to be correlated.
Result
Applying the method to the AD data, it identifies numerous interesting protein peptides, brain regions, and protein‐structure‐memory paths, which are in accordance with and also supplement existing findings of AD research.
Conclusion
The proposed approach is designed to integrate multiview data via a mediation framework. Under biological mechanistic assumptions, it articulates the underlying mechanisms and interrelations between data modalities and further infers disease pathology.