Abstract
Estimating age at death narrows the pool of potential donors in mass disasters and criminal investigations. In this study, we developed a capillary electrophoresis-based SNaPshot assay for age prediction of costal cartilage and used it to analyze DNA methylation at 11 CpG sites across six genes in 136 samples from deceased Koreans aged 28-84 years. To develop the predictive model, DNA methylation levels at these sites from a training set of 83 samples were analyzed using multivariate linear regression in five ways. We then compared the performance parameters calculated from the training set and a test set of 53 samples. Considering experimental simplicity, we selected a model that incorporates four CpGs (MIR29B2CHG_C2, FHL2_C4, TRIM59_C3, and KLF14_C3) as the optimal age prediction model, demonstrating high performance with a mean absolute error of 4.60 years and a root mean square error of 5.41 years in the test set. Subsequently, we developed a multiplex SNaPshot system covering CpGs included in the optimal model, requiring a minimum of 4 ng of bisulfite-converted DNA for reliable prediction and demonstrating multi-tissue applicability, particularly in blood and buccal swabs. We believe this tool will support forensic investigations, including the identification of victims and missing persons.