Integrative analysis of proteomic and MRI data reveals protein associations with brain imaging features in Alzheimer’s disease

Abstract

Background: Alzheimer’s disease (AD) is a progressive neurodegenerative disorder and the leading cause of dementia worldwide. Despite extensive research, the complex molecular mechanisms underlying AD remain incompletely understood, limiting diagnostic and therapeutic advancements. Research design and methods: We presented an integrative, multi-layer computational framework to highlight proteins associated with AD-related brain changes using imaging, proteomic, genetic, and network-based analyses. Utilizing data from the Alzheimer’s Disease Neuroimaging Initiative (ADNI), we combined cerebrospinal fluid (CSF) proteomics with structural brain MRI features through a supervised multi-omics integration method. Results: This approach enabled the identification of key proteins linked to imaging traits. To contextualize these findings, proteins were mapped to their corresponding genes, investigated AD brain-imaging genetic associations through genome-wide association studies (GWAS) and applied network-based analyses. Proteins highlighted from both analyses were further verified in brain-specific databases to assess their functional roles, recurrence across studies, and spatial expression. Five proteins–APP, VGF, APOE, SCG3, and NCAN–were consistently associated with imaging-derived traits and are implicated in neurodegenerative mechanisms. Conclusions: This study highlights the critical role of integrating imaging and proteomic data as part of the genotype-to-phenotype roadmap for AD, revealing molecular underpinnings of brain changes and offering a blueprint for the development of targeted therapeutic strategies.