Son, D. R.; Loh, E. Y.- H.; Jeong, H.; Ma, J.; Eichler, E. E.; Yi, S. V.

Telomere-to-telomere (T2T) genome assemblies resolve DNA methylation of complete ape genomes, including repetitive and satellite-rich compartments that were largely inaccessible in previous primate reference genomes. Here, integrating complete ape genome assemblies with long- and short-read germline and somatic DNA methylomes, we demonstrate dual roles of germline DNA methylation. While solidifying germline DNA methylation as a major driver of genome-wide long-term CpG erosion, we report an unexpected reservoir of hypomethylated CpGs that exist in previously inaccessible regions. Specifically, peri/centromeric satellites are markedly hypomethylated in sperm across great apes, despite their heterochromatic context and rapid sequence turnover. Moreover, we found their hypomethylation extends into adjacent non-satellite DNA at the boundaries of satellite-rich heterochromatic sequences and adjacent euchromatic sequences, forming centromeric hypomethylated extension domains, or CHEDs. CHEDs are enriched in recently duplicated genes, and CHED-associated genes show reproducible, enriched expression in the brain and testis. Extending our analyses to other structurally dynamic regions including lineage-specific insertions, recent segmental duplications and structurally divergent regions, we show that these regions are also CpG-rich and relatively hypomethylated in sperm. Together, our results reveal a unique germline hypomethylation landscape in ape genomes in which structurally dynamic regions act not only as substrates of rapid genome evolution, but also as transient reservoirs of CpG-rich sequence contexts promoting evolutionary innovations involving new genes and tissue-biased expression.