Blood-based epigenome-wide analyses on the prevalence and incidence of nineteen common disease states

Abstract

BackgroundBlood DNA methylation can inform us about the biological mechanisms that underlie common disease states. Previous epigenome–wide analyses of common diseases often focus solely on the prevalence or incidence of individual conditions and rely on small sample sizes, which may limit power to discover disease–associated loci.ResultsWe conduct blood–based epigenome-wide association studies on the prevalence of 14 common disease states in Generation Scotland (nindividuals≤18,413, nCpGs=752,722). We also utilise health record linkage to perform epigenome–wide analyses on the incidence of 19 disease states. We present a structured literature review on existing epigenome–wide analyses for all 19 disease states to assess the degree of replication within the existing literature and the novelty of the present findings. We identify 69 associations between CpGs and the prevalence of four disease states at baseline, of which 58 are novel. We also uncover 64 CpGs that associate with the incidence of two disease states (COPD and type 2 diabetes), of which 56 are novel. These associations were independent from common lifestyle risk factors. We highlight poor replication across the existing literature. Here, replication was defined by the reporting of at least one common gene in >2 studies examining the same disease state. Existing blood–based epigenome–wide analyses showed evidence of replication for only 4/19 disease states (with up–to–15% of unique genes replicated for lung cancer).ConclusionsOur summary data and structured review of the literature provide an important platform to guide future studies that examine the role of blood DNA methylation in complex disease states.

Discovery of shared epigenetic pathways across human phenotypes

“…We extend previous studies both in terms of methodology and in the phenotypes studied. The variance estimates obtained for BMI agree with previous estimates 5 , as does our finding of a strong negative correlation of epigenetic effects between BMI and ratio of high density lipoprotein over total cholesterol 4 . We highlight novel CpG covariances among hypertension and osteoarthritis, ratio of high density lipoprotein over total cholesterol and osteoarthritis, BMI and hypertension, and BMI and asthma, and our results imply that methylation patterns of cholesterol metabolism related genes in whole blood are associated with osteoarthritis pathogenesis and that there is potentially a complex, yet to be fully explored, relationship between hypertension and asthma.…”

“…Epigenetic mechanisms influence gene expression, cell differentiation, tissue development, and disease susceptibility 1,2,3 . Measuring and tracking epigenetic changes through disease progression can provide insight into disease pathogenesis 4 , elucidate environmental and lifestyle factors influencing health, and provide biomarkers for disease diagnosis and risk stratification 5 . To date, most studies have focused on determining the epigenetic basis of traits individually.…”

Antidepressant Exposure and DNA Methylation: Insights from a Methylome-Wide Association Study

“…Seven of the CpGs significantly hypermethylated with antidepressant use have been reported previously to also be significantly hypermethylated with incident and/or prevalent type 2 diabetes (T2D) 59 in GS. Previous epidemiological studies have indicated that antidepressant use leads to an increased risk of T2D onset in a time- and dose-dependent manner 60,61 .…”

“…59 , Sex 71 , Age 72 , Alzheimer's disease Braak stage73 , Estimated glomerular filtration rate 69 , Schizophrenia 70 .0039 6.46x10 -8 0.027 0.0052 3.0x10 -7 C-reactive protein levels (basic model)62 , Chronic pain (basic model)59 , Type 2 diabetes Antidepressant Exposure and DNA Methylation 5 (basic model) 59 , Sex 71 , Age 72 cg02183564 7 76874892 CCDC146 0.019 0.0028 3.59x10 -11 0.019 0.0037 5.17x10 -7 Birthweight 72 , C-reactive protein levels (basic model)62 , Gestational age, Chronic pain (basic model)59 , Type 2 diabetes (basic model)59 , .0032 9.84x10 -9 0.020 0.0042 3.3x10 -6 C-reactive protein levels (basic model)62 , Incident chronic pain (basic model)59 , .0026 5.56x10 -8 0.019 0.0035 3.85x10 -8 C-reactive protein levels (basic model)62 , Incident type 2 diabetes (basic model)59 , Chronic kidney disease (basic & fully adjusted model)59 , Chronic pain (basic model)…”

Feature pre-selection for the development of epigenetic biomarkers

“…Numerous studies have shown that levels of DNA methylation (DNAm) at various CpG sites can correlate with health-related traits, such as body mass index (BMI), smoking status [1], and incident diseases [2, 3, 4]. DNAm is an epigenetic modification whereby methyl groups are dynamically attached and removed at various genomic positions (often on the cytosine of a C-G dinucleotide; CpG) throughout an individual’s lifetime.…”