MskAge - An Epigenetic Biomarker of Musculoskeletal Age Derived from a Genetic Algorithm Islands Model

BackgroundAge is a significant risk factor for functional decline and disease of the musculoskeletal system, yet few biomarkers exist to facilitate ageing research in musculoskeletal tissues. Multivariate models based on DNA methylation, termed epigenetic clocks, have shown promise as markers of biological age. However, the accuracy of existing epigenetic clocks in musculoskeletal tissues are no more, and often less accurate than a randomly sampled baseline model. ResultsWe developed a highly accurate epigenetic clock, Msk-Age, that is specific to tissues and cells of the musculoskeletal system. MskAge was built using a penalised genetic algorithm islands model that addresses multi-tissue clock bias. The final model was trained on the transformed principal components of CpGs selected by the genetic algorithm, which are significantly enriched for pathways terms related to the skeletal system and mesenchyme development. We show that MskAge tracks epigenetic ageing ex vivo and in vitro. Epigenetic age estimates are rejuvenated to zero with cellular reprogramming and are accelerated at a rate of 0.45 years per population doubling. Remarkably, MskAge explains more variance associated with in vitro ageing of fibroblasts than the purpose-developed skin and blood clock. ConclusionThe precision of MskAge and its ability to capture perturbations in biological ageing make it a promising research tool for musculoskeletal and ageing biologists.