Figure 1. Fluctuating CpG sites (fCpG) and modulators of tumor age.

(A) Balanced fCpGs stochastically oscillate between the unmethylated, hemi-methylated, and methylated states. (B) During tumor cell division, methylation and demethylation events occur stochastically. As illustrated for the first three generations of an exponential tumor expansion, the average methylation status ($\beta$-value) evolves over time. (C) In silico modeling of the fCpG dynamics, based on a simple birth-death model of tumorigenesis. The average methylation value ($\beta$) is simulated for three independent fCpGs, whose initial states in the first tumor cell are fully methylated (red), hemi-methylated (yellow), and unmethylated (blue), respectively. Insert: the number of tumor cells over time. Simulation parameters: cell proliferation rate $\alpha =0.17$ divisions/day; cell death rate $\lambda =0.15$ deaths/day; (de-)methylation rate $\mu =0.002$ flips/division. See Methods for details. (D) The age of a tumor at detection is modulated by relative strengths of intrinsic proliferation and growth-suppression induced by the immune tumor microenvironment (TME); tumors will reach a detectable size faster when proliferation is greater and/or the TME is weaker.