A minimal number CpGs of ELOVL2 gene for a chronological age estimation using pyrosequencing

• Published in: Forensic science international 2021-01, Vol.318, p.110631-110631, Article 110631
• Main Authors: Sukawutthiya, Poonyapat, Sathirapatya, Tikumphorn, Vongpaisarnsin, Kornkiat
• Format: Article
• Language: English
• Subjects: Accuracy; Age determination; Age groups; Blood; Chronological age; Chronology; Correlation analysis; CpG islands; Deoxyribonucleic acid; DNA; DNA methylation; ELOVL2; Epigenetics; Forensic; Forensic science; Forensic sciences; Genes; Genomes; Markers; Postmortem sample; Prediction models; Probability distribution; Pyrosequencing; Regression analysis; Root-mean-square errors; Sample size; Statistical analysis
• ISSN: 0379-0738; 1872-6283
• DOI: 10.1016/j.forsciint.2020.110631

•The methylation of ELOVL2 gene is associated to human chronological age.•Linear correlation between age and % methylation of 2 CpGs of ELOVL2.•The least prediction model’s error of 5.59 years in the age group over 20 years.•Productive to the cost and time with a high-performance test’s value. Chronological age estimation is an important piece of human identification used in forensic practice. Epigenetic modifications, especially DNA methylation, have been proposed to predict age. The methylation of the ELOVL2 gene is one of the age-related markers that could be tested in fresh or postmortem blood sample. We study the use of DNA methylation markers on the ELOVL2 gene and develop a prediction model to estimate the age from a postmortem blood sample using pyrosequencing. From 100 anonymous blood samples, a correlation study of DNA methylation and age was investigated. The regression analysis revealed 2 CpG sites for model prediction with an adjusted R2 value of 0.7 (p < 0.01). The model explained 74% of the variation in postmortem blood samples (n = 36) with a prediction error (RMSE) of 10.2 years and a mean absolute deviation (MAD) of 7.1 years, whereas the model (excluding a younger age group) had improved with a RMSE of 5.6 years and a MAD of 4.2 years. The performance parameters were analyzed in several simulated models and indicated that these markers are advantageous for age estimation in forensic scenarios. Finally, a robustness and reproducibility of the pyrosequencing technique would enable this approach to be the part of an age prediction in forensic investigation.