DNA methylation signatures for chromatinopathies: current challenges and future applications

Abstract

Pathogenic variants in genes that encode epigenetic regulators are the cause for more than 100 rare neurodevelopmental syndromes also termed "chromatinopathies". DNA methylation signatures, syndrome-specific patterns of DNA methylation alterations, serve as both a research avenue for elucidating disease pathophysiology and a clinical diagnostic tool. The latter is well established, especially for the classification of variants of uncertain significance (VUS). In this perspective, we describe the seminal DNA methylation signature research in chromatinopathies; the complex relationships between genotype, phenotype and DNA methylation, and the future applications of DNA methylation signatures.

Fig. 1

Generating a new DNA methylation signature for a specific chromatinopathy, begins with obtaining DNA samples from individuals clinically diagnosed with that specific chromatinopathy and control individuals. The resulting DNA methylation signature is then validated in an independent cohort of individuals with the same chromatinopathy, which provides a sensitivity estimate. Specificity of the signature is then estimated by analyzing methylation at signature sites in a large independent cohort of controls and cohorts of individuals diagnosed with other chromatinopathies. In the bottom plot, the KBG syndrome signature demonstrates high sensitivity and specificity, as the KBG validation cohort are classified as “case-like” with high prediction scores, whereas control individuals (validation set) and individuals with other diagnoses, Kabuki and Weaver syndromes, are classified as “control-like” with low prediction scores. As with all signatures, a classification of “control-like” means an individual does not have a casual variant for that specific chromatinopathy; it does not mean that the individual has no diagnosable chromatinopathies. To determine an individual’s epigenotype at signature sites, we utilize a combination of principal component analysis, hierarchal clustering, and machine learning models. Once a DNA methylation signature is generated for a chromatinopathy, it is annotated to identify the location of sites in the genome including overlapping genes and specific location within gene structure (promoter, TSS, body, UTR). Pathway enrichment and gene ontology can elucidate which biological processes and molecular functions genes underlying signature sites are involved in. Comparison to other DNA methylation signatures for chromatinopathies can potentially identify common methylation changes and gene targets relevant to exploring therapeutics in the future