Second-Generation Epigenetic Clocks
First-generation epigenetic clocks are useful in the study of phenotypes and healthy aging, but may not select health informative CpG sites because they were trained exclusively on CA (rather than variables more tightly linked to health status12. Second-generation epigenetic clocks are trained on proxy variables (e.g., five plasma proteins, and smoking status) for biological aging in addition to CA. These clocks aim to improve on the performance of the first-generation clocks in predicting disease development and mortality. PhenoAge66 and GrimAge76(Table 3 ) are second-generation clocks and predict mortality more accurately than first generation clocks. However, there are several key limitations. The clinical markers used to generate a composite “biological age” value makes these clocks difficult to implement in many contexts, as such detailed health data may not always be available. Clinical variables used in the development of these clocks may be relevant to aging in older adults but may not be informative of biological aging in pediatric samples. Pediatric-specific second-generation clocks may be needed to assess the accuracy of second-generation clocks in predicting childhood phenotypes.