Alzheimer's disease (AD) is a major public health crisis and a national priority area of high significance. There is a growing recognition that AD research must confront the challenge of elucidating the disease mechanisms using multi-omics and systems biology approach, since neurodegeneration and AD are multifactorial that may be attributed to harmful changes at multi-omics levels and at pathway/network level. Massive amounts of omics data such as SNPs, gene expression, methylation and metabolome data have been generated in many AD studies including the Alzheimer's Disease Neuroimaging Initiative (ADNI). Integrative analyses of such heterogeneous multi-omics data as well as imaging data from same individuals offer great promises to deepen the understanding of biological bases and molecular underpinnings of AD. However, such data also present daunting analytical challenges. In particular, there has been very limited work on analysis of multi-omics and imaging data in AD research, and current methods for integrative analysis of multi-omics and imaging data fail to take advantage of existing biological knowledge such as functional genomics and brain connectivity, which is essential for AD, a complex, multifactorial disease. To address these and other challenges, we seek to develop novel statistical methods that are guided by biological knowledge such as functional genomics and brain connectivity to advance analysis of multi-omics data in AD research.
Our specific aims are as follows.
Aim 1 : To develop Bayesian generalized factor analysis (GFA) methods for multi-omics and imaging data with incorporation of biological knowledge which can be used to uncover molecular drivers for AD that are potential targets for treatments.
Aim 2 : To develop Bayesian generalized bi-clustering (GBC) methods for multi-omics and imaging data with incorporation of biological knowledge which can be used to identify AD subtypes.
Aim 3 : To develop Bayesian sparse GFA and GBC methods for multi-omics and imaging data based on tensor decomposition that can account for heterogeneity across multiple cohorts.
Aim 4 : To develop user-friendly, open-source software tools for implementing the methods in Aims 1-3 with the goal of dissemination to broad biomedical research community and perform extensive simulations studies to assess the proposed methods.
Aim 5 : To analyze multi-omics and imaging data in the ADNI and IMAS using the methods and tools developed in Aims 1-4 to address significant biomedical questions in AD research. In particular, we will leverage open- source paradigms and containerization approaches to deliver maintainable, extensible and portable software tools that are easy to deploy to biomedical research community. These practices will strengthen the rigor and reproducibility of scientific research. Our project is expected to address long-term challenges and opportunities in analysis of multi-omics data in AD research. In addition, analysis of the ADNI and IMAS data will provide valuable insights on the underlying molecular underpinnings of AD.
This study seeks to develop innovative statistical methods for integrative analysiss and advance analysis of multi-omics data with imaging data in Alzheimer's Disease (AD) research. It is expected to deepen the understanding of biological bases and molecular underpinnings of AD towards the goal of developing a precision medicine approach for AD.
