Embryonic origins of epigenetic mosaicism in the human colorectal epithelium

Abstract Somatic cells can acquire epigenetic changes throughout life, generating epigenetic mosaicism among cells of the same type. However, the extent, timing and persistence of this somatic epimosaicism remain poorly understood. Here, by integrating whole-methylome and whole-genome sequencing of 406 single-cell-derived colorectal epithelial clones from 16 donors, we show that, by ~60 years of age, a somatic colorectal lineage accumulates ~1.2 million CpG methylation differences from the consensus epiallelic state. Remarkably, we find that the majority of these differences (~60%) are established during early embryogenesis and are subsequently stably propagated across somatic lineages throughout life, rather than representing age-associated CpG epimutations. These early CpG methylation differences frequently clustered and recurred at specific genomic regions, consistent with embryonic lineage-specific commitment to one of several possible epiallelic states rather than stochastic epimutation. We identified 14,402 such embryonic epiallelic commitment (EEC) regions, which were enriched at CpG islands and gene promoters, particularly those of homeobox genes and zinc-finger/KRAB-domain-containing genes, consistent with a role in early developmental regulation. Notably, EECs were often associated with differences in gene expression across the adult colorectal epithelium, suggesting that their functional consequences may persist into adulthood. EECs also accounted for >70% of LINE-1 promoter loss-of-methylation frequently observed in adult colorectal epithelium and enabled more accurate reconstruction of early developmental phylogenies than somatic mutations alone. This principle extended to the DNA methylation landscape underlying X-chromosome inactivation and escape in female cells. Together, our study reveals early embryogenesis as a major source of lifelong somatic epigenetic mosaicism and establishes clonal multi-omics as a high-resolution framework for tracing its origins, persistence and functional consequences.