Conserved Human CpG Dinucleotides Identify Enriched Ageing-related Signals in Developmental Pathways and the Brain

Abstract

The DNA methylome changes with age. This is observed as both random drift, but also consistent alterations within certain genomic loci enabling the construction of precise age predictors. However, the functionality of these ageing-related modifications remains largely undefined. The CpG dinucleotide is the principal sequence target for these epigenetic DNA marks in differentiated cells. Here, for functional insight, we identified ultra-conserved CpGs (ucCpGs, n [~]167k) lacking observed sequence mutation in large-scale human whole genome datasets (>576k). ucCpGs were enriched, as expected, in lowly-methylated CpG dense loci, due to methylated cytosine hypermutability. Additionally, ucCpGs demonstrated pathogenic evidence (CADD, ClinVar), and enrichment in four-fold degenerate sites, as well as within developmental and ageing-related gene families (AP-2, HOX-L, C2H2-ZNF, etc). Extreme ucCpG clusters ([≥]16 ucCpGs/kb) were enriched for brain-expressed genes, as well as developmental and ageing/mortality pathways. ucCpGs also strongly co-located within ageing-related Differentially Methylated Regions (age-DMRs), highlighting Clustered Protocadherin Gamma, as well as HSPA2 and LHFPL4 genes. These findings further support that functional components of DNAm ageing are intertwined with developmental pathways.