An epigenome wide association study of sub‐phenotypes in Parkinson’s disease.

Abstract

AbstractBackgroundParkinson’s disease (PD) is a highly heterogeneous disorder, encompassing a complex spectrum of clinical presentation including motor, sleep, cognitive and neuropsychiatric symptoms. Together the cumulative effects of these symptoms have huge impacts on individual quality of life. Here we aim to investigate genome wide DNA‐methylation in post‐mortem PD brain samples and test for region specific association with common secondary symptoms of sleep disorder, visual hallucinations and dementia.MethodDNA methylation data was generated on the Illumina 450K microarray for a cohort of 142 individual post mortem brain samples from the Frontal Cortex (FC), Caudate Putamen (CP) and Substantia Nigra (SN). Data was quality controlled following standardized, previously published pipelines. Clinical records for each case were assessed for binary status of sleep disorder, visual hallucinations and dementia. Linear models were used to identify differentially methylated positions (DMPs) for each of these outcomes, controlling for age, sex, post‐mortem interval and inferred cell‐type proportions as covariates.ResultThus far, results been only been generated for the dementia phenotype, with one genome wide significant loci passing multiple testing correction in the frontal cortex (p = 5.25e‐8). This DMP, found in the gene FAM175A / ABRAXAS‐1 (cg14699004), is hypomethylated in PD patients with preserved cognition, with comparable higher measures in PD‐Dementia and control samples. This effect appeared to be FC specific, with no DMPs passing genome wide significance in other brain regions tested.ConclusionPreliminary evidence indicates potential sub‐phenotype associated DMPs, with methylation at one site specifically associated with dementia within the FC. Current analysis is focussing on generating reference results across each of the phenotypes and brain regions assessed. Future work will look at Weighted Gene Correlation Network Analysis to identify networks associated with these individual phenotypes.