Abstract Phenotypic plasticity is thought to assist immune evasion, metastasis, and therapeutic resistance in tumor evolution. Epigenetic mechanisms such as DNA methylation provide for a more easily acquired and flexible, yet semi-stable adaptation by the tumor cell than hardwired genetic changes. Cancer-associated gain of DNA methylation at promoter CpG islands or distal regulatory elements is associated with transcriptional silencing, whereas loss of DNA methylation is observed at late-replicating, lamina-attached regions of the genome, tracking with cumulative cell divisions. One of the most well-documented genomic features predisposing to DNA hypermethylation is Polycomb Repressive Complex (PRC) occupancy in precursor cells, consistent with an epigenetic block to cellular differentiation as an early or predisposing event in carcinogenesis. A central question is whether this epigenetic switch from reversible PRC repression to more stable DNA hypermethylation occurs before onset of clonal expansion, or whether it represents an ongoing process during tumor evolution. Clonal expansion is a pivotal characteristic of cancer, and is thought to be initiated by a genetic alteration in a key gatekeeper driver gene. However, not all normal cells appear to be susceptible to malignant transformation following such an event. Multiple lines of evidence suggest that epigenetic heterogeneity among normal cells may affect their cancer-initiating potential. We have developed a tumor processing pipeline to deeply and comprehensively map DNA methylation patterns in primary human colorectal cancer (CRC) and adjacent tissues, while preserving spatial information, enabling inference of the temporal order of molecular alterations. We employ a suite of technologies to map the spatial heterogeneity of DNA methylation patterns in primary human CRC. These include a deep single-cell whole-genome bisulfite sequencing technique that improves genomic coverage several-fold over existing published protocols, providing DNA methylation measurements at the majority of CpGs in the genome in individual tumor cells. We have also calibrated and validated an epigenetic clock that allows us to accurately measure the relative replicative history of individual cells. We infer phylogenetic relationships among subclones, and relate these to their spatial distribution in the primary tumor. We find that some tumors have high proportions of PRC-associated clonal hypermethylation, indicating that these hypermethylation events were already present prior to clonal expansion, possibly contributing to susceptibility to malignant transformation in the cell-of-origin, whereas other tumors appear to have active PRC-associated CpG island hypermethylation ongoing during tumor expansion, generating subclonal heterogeneity, which may influence post-therapeutic recurrence potential. Citation Format: Peter W. Laird, Nathan J. Spix, Hsiao-yun Milliron, Manpreet Kalkat, Emily Eugster, David Sokol, Emily Jung, Paula Nolte, Kelly K. Krzyzanowski, Toshinori Hinoue, Hui Shen. Spatial reconstruction and temporal inference of DNA methylation alterations in primary human colorectal cancer [abstract]. In: Proceedings of the AACR Special Conference in Cancer Research: Translating Cancer Evolution and Data Science: The Next Frontier; 2023 Dec 3-6; Boston, Massachusetts. Philadelphia (PA): AACR; Cancer Res 2024;84(3 Suppl_2):Abstract nr IA023.