DNA methylation of the ELOVL2, FHL2, KLF14, C1orf132/MIR29B2C, and TRIM59 genes for age prediction from blood, saliva, and buccal swab samples. (January 2019)

• Record Type: Journal Article
• Authors: Jung, Sang-Eun; Lim, Seung Min; Hong, Sae Rom; Lee, Eun Hee; Shin, Kyoung-Jin; Lee, Hwan Young
• Is Part Of: Forensic science international. Volume 38(2019)
• Page Start: 1
• Page End: 8
• Publication Date: 2019-01
• Subjects: Age -- DNA methylation -- Blood -- Saliva -- Buccal swab -- Methylation SNaPshot; Forensic genetics -- Periodicals; Génétique légale -- Périodiques; Forensic genetics; Electronic journals; Periodicals; 614.1
• DOI: 10.1016/j.fsigen.2018.09.010
• Languages: English
• ISSNs: 1872-4973

Abstract

Highlights: DNA methylation of the ELOVL2, FHL2, KLF14, C1orf132/MIR29B2C, and TRIM59 genes was investigated in blood, saliva, and buccal swab samples. DNA methylation was measured using a multiplex methylation SNaPshot assay. DNA methylation of the ELOVL2, KLF14 and TRIM59 genes showed strong age-correlation from blood, saliva, and buccal swabs. Blood model showed strong age correlation of r = 0.956 and a high prediction accuracy with a MAD of 3.478 years. Saliva model showed strong age correlation of r = 0.948 and a high prediction accuracy with a MAD of 3.552 years. Buccal swab model showed strong age correlation of r = 0.925 and a high prediction accuracy with a MAD of 4.293 years. A tissue-combined model showed strong age correlation of r = 0.937 and a high prediction accuracy with a MAD of 3.844 years. Abstract: Many studies have reported age-associated DNA methylation changes and age-predictive models in various tissues and body fluids. Although age-associated DNA methylation changes can be tissue-specific, a multi-tissue age predictor that is applicable to various tissues and body fluids with considerable prediction accuracy might be valuable. In this study, DNA methylation at 5 CpG sites from the ELOVL2, FHL2, KLF14, C1orf132/MIR29B2C, and TRIM59 genes were investigated in 448 samples from blood, saliva, and buccal swabs. A multiplex methylation SNaPshot assay was developed to measure DNA methylation simultaneously at the 5 CpG sites. Among the 5 CpG sites, 3 CpG sites in the ELOVL2, KLF14 and TRIM59 genes demonstrated strong correlation between DNA methylation and age in all 3 sample types. Age prediction models built separately for each sample type using the DNA methylation values at the 5 CpG sites showed high prediction accuracy with a Mean Absolute Deviation from the chronological age (MAD) of 3.478 years in blood, 3.552 years in saliva and 4.293 years in buccal swab samples. A tissue-combined model constructed with 300 training samples including 100 samples from each blood, saliva and buccal swab samples demonstrated a very strong correlation between predicted and chronological ages (r = 0.937) and a high prediction accuracy with a MAD of 3.844 years in the 148 independent test set samples of 50 blood, 50 saliva and 48 buccal swab samples. Although more validation might be needed, the tissue-combined model's prediction accuracies in each sample type were very much similar to those obtained from each tissue-specific model. The multiplex methylation SNaPshot assay and the age prediction models in our study would be useful in forensic analysis, which frequently involves DNA from blood, saliva, and buccal swab samples.