ForSpEC: A compact forensic epigenetic age clock for sperm cells with cross‑platform validation

Expanding the use of forensic DNA intelligence requires the development of accurate DNA-based predictive models. The primary challenge in achieving a highly accurate epigenetic age prediction method in sperm lies in the unique biological function of male reproductive cells. In this study, we independently evaluated the Germ Line Age (GLA) calculator, one of the most comprehensive tools for predicting sperm age to date. We then set out to define a streamlined, well-curated set of sperm cell-specific markers suitable for forensic applications, while providing an integrated laboratory tool for targeted methylation profiling using bisulfite amplicon sequencing with Ion AmpliSeq™ technology. DNA was extracted from 212 sperm samples, genome-wide DNA methylation was measured using Infinium MethylationEPIC (EPIC) arrays and methylation beta values were extracted for 260 literature-reported CpG markers. Application of the GLA model to this sperm cell dataset resulted in a mean absolute error (MAE) of 2.82 years. As a next step, predictive analysis based on 147 training set samples identified a subset of eight optimal CpG markers for model building. Due to technical constraints encountered during Ion AmpliSeq™ assay optimization, the final model was refined to include seven CpGs, yielding the ForSpEC (Forensic Sperm Epigenetic Clock) model. ForSpEC predicted age with an MAE of 3.13 years in the 65 test samples. When the EPIC-trained ForSpEC model was applied to Ion AmpliSeq™ data (N = 20), age prediction accuracy remained high, with an MAE of 3.76 years. For comparison, existing predictors were retrained using EPIC data, demonstrating that the seven CpGs comprising ForSpEC rank among the most accurate predictors of sperm epigenetic age. The model was further validated using data from publicly available repositories and the impact of cellular composition on prediction accuracy was additionally addressed by comparing DNA samples isolated from sperm cells with whole semen samples. Overall, this study advances sperm-based epigenetic age estimation by introducing ForSpEC, a highly accurate forensic epigenetic clock that relies on just seven CpG markers. This new method exhibits only a modest reduction in accuracy compared to the GLA model, which relies on 261 CpGs. Moreover, the analyses underscore the high utility of Ion AmpliSeq™ technology for targeted DNA methylation analysis in forensic applications.