Year: 2021
Type: dissertation
Abstract: Age estimation is very useful in forensic sciences for identification purposes. In recent years, DNA methylation (DNAm) has arisen as a promising tool for age prediction. In consequence, age prediction models (APMs) have been developed using many age-correlated genes, different methodologies and several biological samples. The main aim of this study was to develop APMs based on previously known age-associated genes, including ELOVL2, EDARADD, FHL2, PDE4C, C1orf132, KLF14 and TRIM59, using different tissues with forensic interest. In addition, we investigated putative differences in DNAm between blood from living and deceased individuals and according to increasing age, sex, ancestry and postmortem interval (PMI). DNAm levels were evaluated in 267 biological samples (including blood, teeth, bones and buccal swabs) from living and deceased Portuguese individuals through bisulfite polymerase chain reaction (PCR) Sanger sequencing and multiplex SNaPshot methodologies. Additionally, nine skeletonized individuals were tested for DNAm analysis. Linear regression models were used to analyze relationships between DNAm and chronological age. Promising results were obtained in our study. Using SNaPshot, the most accurate APM built in 56 blood samples from living individuals included three CpGs (in genes ELOVL2, FHL2 and C1orf132), showing a Mean Absolute Deviation (MAD) between predicted and chronological ages of 4.25 years; in 59 deceased individuals the most accurate APM was built with four CpGs (in genes ELOVL2, FHL2, C1orf132 and TRIM59) with a MAD of 5.36 years. The Sanger sequencing allowed to build accurate APMs in 53 blood samples from living individuals (MAD = 5.35 years), with the best CpGs at ELOVL2, FHL2, EDARADD and PDE4C, and in 47 blood samples from deceased individuals (MAD = 6.08 years), with the best CpGs at ELOVL2, FHL2, C1orf132, EDARADD and PDE4C. In the combined set of blood samples from living and deceased individuals, the best APMs were built with CpGs at ELOVL2, FHL2, C1orf132 and TRIM59 (MAD = 4.97 years) using SNaPshot, and with CpGs at genes ELOVL2, FHL2, EDARADD and PDE4C (MAD = 6.21 years) through Sanger sequencing. Moreover, our study highlighted the most promising APM in 29 bones from autopsies, using Sanger sequencing, built with six CpGs at ELOVL2, EDARADD and C1orf132, and revealing a MAD of 2.56 years. Through SNaPshot, the most accurate APM (MAD = 7.18 years) in 31 bones from autopsies was built with two CpGs in genes FHL2 and KLF14. Potential useful APMs were developed in the other tissues, including: in 21 bones from Bodies Donated to Science with two CpGs in genes FHL2 and PDE4C (MAD = 4.67 years) using Sanger sequencing; in buccal swabs with TRIM59 CpG using SNaPshot in 35 samples (MAD = 6.73 years) and with two CpGs at ELOVL2 using Sanger sequencing in 23 samples (MAD = 8.32 years); in teeth with two CpGs at ELOVL2 and KLF14 (MAD = 7.07 years) in 23 samples using SNaPshot. Additionally, we developed several multi-tissue APMs, obtaining the best accuracy (MAD = 6.49 years) in a group with bones from autopsies and blood and teeth from living and deceased individuals using SNaPshot (with CpGs at ELOVL2, KLF14 and C1orf132). In this same group, Sanger sequencing allowed the development of an APM with nine CpGs at genes ELOVL2, EDARADD, PDE4C and FHL2 and a similar value of MAD (6.42 years). Ours results revealed that the model accuracy depends on the chronological age of individuals, in concordance with previous studies, and that there is no influence of sex and PMI in DNAm levels. The possibility that age markers might be population specific and that postmortem changes can alter the methylation status among specific loci was also suggested by our data. DNAm changes were observed due to treatment of the body after death. In regards to bones from skeletonized individuals, the amount of extracted DNA was not sufficient for a subsequent successful DNAm age estimation. Moreover, it has been observed that the APMs developed in blood from living individuals cannot be applied to deceased individuals with a similar MAD value. However, the APMs developed in blood from deceased individuals when applied to living individuals revealed a similar MAD. This can be a relevant issue in future application of the developed APMs in forensic contexts. In conclusion, our study is a relevant contribution for DNAm age research, showing the usefulness of both bisulfite PCR sequencing and multiplex SNaPshot methodologies for forensic analysis in several tissue types. The herein developed APMs seem to be informative and could have potential applications in different forensic contexts. Despite of this, our study could be improved with the use of larger sample sets, with higher statistical power, and being more representative of age range and different tissue types.
Author: Maria Helena Dias
Language: English
Subfield: Genetics
Domain: Life Sciences
Open Access status: green
OpenAlex ID: W3211593685