Fig. 1
From: A comparison of epigenetic mitotic-like clocks for cancer risk prediction
The epiTOC2 model. a epiTOC2 is based on a DNAm transmission model between cell-generations, from which a mathematical expression can be derived that relates the measured DNAm fraction at a given CpG site i at cell division time t, βi(t), to the site’s de novo methylation probability δi and ground-state (i.e., fetal stage) methylation βi0, as well as the chronological age of the individual and the intrinsic annual rate of stem cell division per stem cell (IR), which is CpG independent. These parameters can be estimated using a large cohort of healthy individuals for which DNAm at specific CpG sites (the epiTOC2-CpGs) has been measured in a common tissue (we do this for blood). An example of an epiTOC2 CpG demonstrating a reasonably good fit to the model is shown together with the estimated parameters. The 3-dimensional bar chart displays the number of epiTOC2-CpGs from a total of 163 (z-axis, count) with particular estimated de novo methylation probabilities (x-axis, delta δ) and intrinsic stem cell division rate estimates (y-axis, IR), as derived by fitting the epiTOC2 model to 656 whole blood samples. The bar chart reveals a clear mode at IR = 35, which we take as an estimate of the intrinsic rate of stem cell division of blood. b With the δi and βi0 parameters estimated in a, which are assumed to be approximately tissue independent, one can estimate the total number of stem cell divisions in any given tissue x of an individual s, TNSC(s,x), from the measured DNAm fraction over the 163 epiTOC2 CpGs. If the age of the person is known, one can estimate the average lifetime rate of stem cell division per stem cell of the person, and for large cohorts of healthy individuals, the intrinsic rate of tissue type x can be estimated by taking the average or median over all samples s