New computational tools for understanding and predicting AD via age-associated DNA methylation changes

Funding: 1RF1NS128145  PI: Wang L

Alzheimer’s disease (AD) is the most common neurodegenerative disorder, with late-onset AD affecting about 1 in 9 people over 65 years old in the US. The increasing elderly population in the US makes AD a major public health concern and one of the most financially costly diseases. Currently, a major challenge is the lack of reliable, minimally invasive, inexpensive biomarkers to aid the diagnosis, prognosis, and ultimately development of new AD treatment strategies. One potential source of biomarkers for AD is DNA methylation (DNAm). Changes in DNAm have been implicated in both aging and AD. Moreover, DNAm is relatively stable and can be easily detected. DNAm is an epigenetic mechanism at the interface of the genome and environment, and it is influenced by aging and many lifestyle factors such as smoking, diet, and exercise, which in turn might modify the risk of AD.

In the past few years, we have developed several innovative open-source software for DNAm and other genomics data analyses. In this proposal, building on our previous experiences in developing and applying tools for integrative genomics analyses of large-scale heterogeneous datasets, we propose to harmonize a large number of DNAm datasets to clarify the role of DNAm in aging and AD, to develop a web interface that disseminates the analyses results, and to develop epigenetic clocks tailored for predicting AD phenotypes. We hypothesize a number of DNAm-based regulatory changes are relevant to both aging and AD, and some age-associated DNAm changes also contribute to AD onset and progression.

In Aim 1, we will aggregate, harmonize, and meta-analyze a large number of DNAm aging datasets measured in brain and blood samples to identify DNAm changes associated with aging and AD, and determine age-associated DNAm differences that also contribute to AD. We will develop two tools: (1) a searchable web interface that clarifies the role of DNA methylation in aging and AD and (2) an open-source R package for performing meta-analyses of DNAm methylation regions. In Aim 2, we will develop a new epigenetic clock tailored for predicting AD phenotypes. The diagnostic and prognostic values of the new epigenetic clock will be evaluated using available CSF biomarkers and clinical cognitive outcomes and compared with known clinical and genetic factors, as well as currently available plasma biomarkers.

The searchable web interface will significantly enhance our understanding and enable new biological insights on the role of age-associated epigenetic changes in AD. The new epigenetic clock tailored to predicting AD phenotypes will facilitate the development of surrogate biomarkers that provide a degree of objectivity for monitoring disease progression in clinical trials,  as well as assessing individualized risk profiles for AD diagnosis and prognosis. The successful completion of the project will also provide us with computational pipelines and tools that can be easily adapted and applied to analyze datasets generated for other types of dementias.