Fig. 6

Proposed model for the influence of blood cell proportions on age prediction by epigenetic clocks. Three hypothetical people are highlighted. Person A is young and has a blood cell composition consisting of mostly naive immune cells, leading to epigenetic clocks rightfully predicting them to be young. Conversely, Person B is old and has a relatively high proportion of activated immune cells in their blood, causing the epigenetic clocks to estimate them to be old. For Person C, who has the same chronological age as Person B but a blood cell composition closer resembling that of a young person, epigenetic clocks would likely underestimate their age. Note that our data indicate that the prediction errors of the various clocks (e.g., due to measurement error of the arrays) also contribute to deviations between chronological and predicted age and hence biological phenomena do not explain deviations in full