Over the past decade, these GEITP pages have covered many genome-wide association studies (GWAS), because these projects are attempts to detect DNA sequence variants that are (statistically significantly) associated with complex diseases (e.g. type-2 diabetes, schizophrenia, obesity, cancer) or responses to drugs or environmental toxicants (e.g. efficacy, therapeutic failure, toxicity) or quantitative traits (e.g. height, body mass index, IQ). And GWAS results have indeed identified thousands of genetic variants that are associated with these various multiplex phenotypes. However, the single-nucleotide polymorphisms (SNPs) or variants (SNVs) –– that have been identified –– only explain a small fraction of the risk of disease, or drug or toxicant effect, or quantitative trait; we now realize that most of the genetic contribution to human multifactorial traits still remains unidentified. This has been termed “missing heritability.”
It seems likely that other molecular factors (e.g. epigenetic effects and protein biomarkers) are associated with these multifactorial traits. Whereas biomarkers are often considered being “markers of disease”, epigenetic factors are suggested to have a causal effect on disease development (and, most likely, other complex and quantitative traits). Epigenome-wide association studies (EWAS) have therefore been performed –– attempting to identify associations between epigenetic modifications and complex traits; for example, differential epigenetic patterns have been identified (e.g. for asthma, obesity and myocardial infarction). However, it remains unclear whether epigenetic variation is causal in the pathogenesis of disease.
Epigenetic variation can be influenced by genetic variation, binding of transcription factors, or by environmental factors such as smoking. Transgenerational heritability of epigenetic alterations (i.e. epigenetic changes that escape reprogramming during formation of gametes and during embryogenesis) has not been demonstrated in humans. On the other hand, intergenerational heritability has (e.g. a fetus exposed to a certain environment before birth, causing epigenetic changes during this time). Newborns of mothers who have smoked during pregnancy exhibit epigenetic changes that reflect increased activity of specific genes (including AHRR encoding aryl-hydrocarbon receptor repressor) that are involved in metabolism of toxic components in tobacco smoke; these changes in AHRR are similar to those observed among adult smokers. However, activity of these genes is not likely to cause increased risk of smoke-associated disorders; rather, exposure to toxic compounds during embryogenesis could serve as a causal factor for increased disease risk (e.g. variation in AHRR DNA-methylation).
As these GEITP pages have stated many times, epigenetic effects include DNA-methylation, RNA-interference, histone modifications, and chromatin remodeling. Assays are now available for the first two, but research is still being refined to create assays for the latter two. The current article is about DNA-methylation. Authors [see attached article] combined SNV data and DNA-methylation data with measurements of protein biomarkers for cancer, inflammation or cardiovascular disease –– to investigate the relative contribution of genetic and epigenetic variation on biomarker levels (121 protein biomarkers were analyzed, relative to DNA-methylation at 470,000 genomic positions, and to more than 10 million SNVs). GWAS and EWAS were performed, using between 698 and 1,033 samples (depending on data availability for the different phenotypes). Most GWAS loci were cis (i.e. near or inside the gene) regulatory, whereas most EWAS loci were located in trans (i.e. far away from the nearest gene). All EWAS signals that overlapped with a GWAS locus were driven by underlying genetic variants, and three EWAS signals were confounded by smoking.
While some cis-regulatory SNVs for biomarkers appeared also to have an effect on DNA-methylation levels, cis-regulatory SNVs for DNA-methylation were not found to affect biomarker levels. Associations between protein biomarker and DNA-methylation levels were seen at numerous loci throughout the genome. Authors conclude that these associations likely reflect underlying patterns of genetic variants, or specific environmental exposures, or these patterns might represent secondary effects to the pathogenesis of disease. Without a doubt, many future (large cohort) EWAS are predicted. 🙂
DwN
PLoS Genet Sept 2o17: 13: e1007005