Fig. 6. FMC dynamics can further be observed in chronic and acute leukemia.
a, The variance of the fCpG methylation distribution experiences a gradual increase with age in healthy individuals. The confidence band was calculated via bootstrapping and represents 95% confidence intervals. b, Left: the variance (center line, median; box limits, upper and lower quartiles; whiskers, 1.5 interquartile range) of paired blood samples taken 10 years apart (1997 and 2007) also exhibits a small but marked increase (0.37 Cohen’s d, P = 2.8 × 10−5 two-sided paired t-test). Right: a scatterplot showing the matched variance per individual sample taken 10 years apart, demonstrating that the variance typically rises with age. ***P < 0.001. c, The variance of the fCpG methylation distribution (center line, median; box limits, upper and lower quartiles; whiskers, 1.5 interquartile range) is a proxy for the rapidity of the clonal expansion within the blood. In normal samples, the large stem cell population size leads to the methylation distribution being concentrated near 50% (as one would expect for uncorrelated oscillators). However, as a clonal cancerous population expands, clonal peaks begin to separate from the 50% peak. In the case of acute lymphoblastic leukemia (ALL), the large, well-separated peaks near 0% and 100% are indicative of a single clonal population making up the majority of the remaining stem cells following rapid growth; CMML, chronic myelomonocytic leukemia; MDS, myelodysplastic syndrome; ET, essential thrombocythemia; PRV, polycythemia vera; PMF, primary myelofibrosis; CML, chronic myeloid leukemia; AML, acute myeloid leukemias. d, Simulations confirm that a simple model of hematopoiesis can recapitulate the observed methylation distribution overserved in human data. Data represent mean ± s.e.m.; CHIP, clonal hematopoiesis of indeterminate potential.