DNA METHYLATION DYNAMICS IN ALZHEIMER’S DISEASE DIAGNOSIS AND PROGRESSION

The appearance of clinical symptoms in Alzheimer's disease (AD) trails molecular changes in the brain, hindering effective treatment. Reliable early molecular diagnosis has been a challenge- current methods are invasive, time-consuming, and/or expensive. Several studies have shown that AD etiology and progression may be linked to altered DNA methylation (DNAm) in genes associated with AD pathology (e.g.ABCA7). To identify peripheral blood (PB) biomarkers of prodromal AD, we examined DNAm changes in an Alzheimer's Disease Neuroimaging Initiative (ADNI) cohort. DNAm was analyzed on Illumina EPIC chips in 649 individuals categorized as healthy, mild cognitive impairment (MCI), or AD. Longitudinal DNA samples (baseline, +1, +2 years) were obtained from all subjects, with additional samples from subjects converting from healthy to MCI, healthy to AD, or MCI to AD. Samples were randomized using a modified incomplete balanced block design, whereby all samples from a subject were on the same chip, with remaining chip space occupied by age- and sex-matched samples from a subject with a different diagnosis. Unused chip space was leveraged for technical reproducibility assessment via replicated DNA samples. A total of 1920 samples were analyzed, including 1719 unique samples and 201 replicates. One sample failed the run and four additional samples failed quality control since ≥1 % of CpG sites had a detection p-value >0.05. Average CpG call rate was 864,640. In paired replicates, the mean within-subject Pearson correlation was 99.42%(SD=1.2%), 95% confidence interval of 97.99%-99.84%. Replicates on the same plate with the same scan date had a mean correlation of 99.63%(SD=0.32%), whereas replicates that didn't share either of these attributes had a mean correlation of 99.25%(SD=1.8%). All correlations reported are based on raw beta-values, and are not a function of normalization. Analysis is currently underway to identify potential DNAm changes associated with disease state and progression. Our work is the largest study thus far to investigate DNAm dynamics that associate with disease progression, serving as potential biomarkers or mechanistic indicators. The integration of this data with the well-characterized ADNI dataset will leverage the rich phenotypic and genotypic information to examine for epigenetic variants that may alter disease course and/or diagnosis.