Epigenetic Age Signatures in Postmortem Rib Samples

Abstract

Skeletal remains, often partially or completely decomposed, are among the most common biological forensic samples found at crime scenes. Analyzing these incomplete specimens to estimate the age of the deceased is crucial. Previous studies on DNA methylation-based age prediction in bones have not evaluated differences across skeletal elements or clarified how bone type influences prediction accuracy. This study focuses on postmortem rib samples—a common forensic specimen—to develop a DNA methylation-based age prediction model specific to ribs. DNA methylation levels at eight CpG sites within the ELOVL2, FHL2, KLF14, and FAM123C genes were analyzed using pyrosequencing in 81 postmortem rib samples and 112 postmortem blood samples, with 50 individuals providing both sample types simultaneously. The rib-derived age prediction model exhibited an R2 value of 0.908, whereas the blood model achieved an R2 value of 0.927. For the rib model, the mean absolute deviation (MAD) of the training set was 4.813 years, and the MAD of the testing set was 5.084 years. The blood model showed slightly higher accuracy in predicting the age of the same individuals. Notably, cross-tissue application of models led to significant prediction bias, emphasizing the necessity of tissue-specific calibration for methylation-based age estimation. Exploratory analysis of postmortem sternum, rib, and frontal bone samples from 12 individuals revealed no statistically significant differences in methylation levels or age estimates across bone types. However, broader generalizability of the rib model to these skeletal elements requires validation in larger, independent cohorts. This work establishes a robust age prediction framework for rib samples, highlights the critical role of tissue specificity in epigenetic forensic models, and provides preliminary evidence for potential cross-bone applicability.