Answers are grounded in this paper and cite their sources.
DNA methylation establishes and maintains cellular identity, is copied at every cell division, and changes during ageing and in cancer. Yet how it evolves within individual somatic lineages over the human lifespan has not been observable. We reconstruct the evolutionary history of the methylome in human blood from matched whole-methylome and whole-genome sequencing of 1,183 single-cell derived haematopoietic colonies from 13 individuals, spanning 17 weeks post conception to 77 years of age, healthy haematopoiesis to malignancy, and including monozygotic twins. Heritable methylation change exceeds the rate of single base substitution by six orders of magnitude before birth and four orders thereafter, often affecting whole tracts of CpGs at once. Each blood cell carries ~300,000 somatic methylation changes before birth, including >3000 at large segments that remain stable for decades. We term these large segments embryonic epialleles: CpG tracts at which lineages of the same cell type within an individual carry different heritable methylation states, of which we identify ~250,000 across the cohort, with many recurring between individuals. Against this developmental mosaic, we time X-chromosome inactivation to the first few divisions of life. We show that epigenetic age reports the division history of whichever clones dominate blood, rather than the age of the individual, reframing the interpretation of methylation biomarkers. In clonal expansions and blood cancer, many methylation differences between an expansion and its background prove to be embryonic states inherited by the founder cell. Expansions do, however, converge on methylation change at Polycomb-associated regulatory regions, gaining methylation in most clades but losing it where DNMT3A is mutated. A few changes, at established cancer genes, including CDKN2B, ABL1 (in BCR::ABL1) and CACNA1D, arise only within clonal expansions, and bear the hallmarks of epigenetic drivers. Methylation states determine which transcript a gene produces, and whether it is expressed at all. Together with their heritability and variability between blood cells, these phenotypic consequences establish methylation mosaicism as a pervasive additional substrate for Darwinian somatic evolution.
No takes yet. Share an insight, caveat, or question.
Nangalia et al. (2026) studied this question.
Synapse has enriched 8 closely related papers on similar clinical questions. Consider them for comparative context: