Chronological age estimation from DNA methylation data in a Serbian population: a SNaPshot-based forensic model

Abstract

Accurate chronological age estimation from biological material represents a valuable investigative tool in forensic genetics, particularly in cases involving unidentified persons or where no DNA database match exists. This study aimed to develop and validate the first DNA methylation–based age prediction model tailored to the Serbian population. Methylation fractions at 40 CpG sites located within seven age-associated gene loci (ELOVL2, FHL2, KLF14, TRIM59, MIR29B2CHG, PDE4C, and EDARADD) were quantified using the SNaPshot assay across 188 peripheral blood samples collected from healthy volunteers aged 18–70 years. The dataset was divided into a training set (N = 142) and a test set (N = 40), with an additional six samples from simulated crime scene conditions. Four age prediction models were constructed using multiple regression and machine learning approaches, all demonstrating high predictive accuracy (MAE = 2.04–2.8 years in training; 3.1–3.24 years in test datasets). All models showed notably high accuracy in the youngest age group (18–30 years), whereas prediction precision decreased slightly with age. Importantly, the models maintained robustness when tested on DNA extracted from bloodstains deposited under non-laboratory, uncontrolled conditions, underscoring their forensic applicability.