A study identifies 88 CpG sites in or near 80 genes in saliva samples of 34 identical twin pairs aged 21–55 years, where methylation levels are significantly correlated with age. These sites were validated in a general population sample of 31 males and 29 females aged 18–70 years. Methylation levels of three genes (EDARADD, TOM1L1, NPTX2) show a linear relationship with age over five decades. A regression model using two cytosines from these loci explained 73% of age variance, with an average accuracy of 5.2 years. This model could estimate age from biological samples, aiding forensic science and predicting age-related disease risk in medical screening. The study also found that these genes are enriched for those involved in cardiovascular, neurological, and genetic diseases. The model was validated in three sample sets, showing strong correlation with age. The model's accuracy remained high even when excluding twin data. The study highlights the potential of epigenetic markers for predicting age and disease risk, offering a tool for personalized medicine based on "bio-age" rather than chronological age. The findings suggest that critical regulatory regions of the genome remain under strict epigenetic control, while non-coding regions may show random drift with age. The study provides a framework for future research on epigenetic aging and disease.A study identifies 88 CpG sites in or near 80 genes in saliva samples of 34 identical twin pairs aged 21–55 years, where methylation levels are significantly correlated with age. These sites were validated in a general population sample of 31 males and 29 females aged 18–70 years. Methylation levels of three genes (EDARADD, TOM1L1, NPTX2) show a linear relationship with age over five decades. A regression model using two cytosines from these loci explained 73% of age variance, with an average accuracy of 5.2 years. This model could estimate age from biological samples, aiding forensic science and predicting age-related disease risk in medical screening. The study also found that these genes are enriched for those involved in cardiovascular, neurological, and genetic diseases. The model was validated in three sample sets, showing strong correlation with age. The model's accuracy remained high even when excluding twin data. The study highlights the potential of epigenetic markers for predicting age and disease risk, offering a tool for personalized medicine based on "bio-age" rather than chronological age. The findings suggest that critical regulatory regions of the genome remain under strict epigenetic control, while non-coding regions may show random drift with age. The study provides a framework for future research on epigenetic aging and disease.