Thesis

Forensic age estimation using DNA methylation analysis

Creator
Rights statement
Awarding institution
  • University of Strathclyde
Date of award
  • 2020
Thesis identifier
  • T15469
Person Identifier (Local)
  • 201590047
Qualification Level
Qualification Name
Department, School or Faculty
Abstract
  • Over the last decade, researchers have identified age-related (AR) DNAmethylation (DNAm) markers, which have outperformed all other known AR biomarkers in estimating the chronological age of individuals with high accuracy,using various tissue types. Their accuracy in age estimation has led to them being suggested as a source of intelligence for forensic investigations, to determine the age of unidentified donors of biological samples left at crime scenes. Initially, in this research, different statistical methods have been tested in order to demonstrate which one of them is optimum for the identification of AR CpG sites.The selected method was then used to identify saliva specific AR CpG markers using DNAm profiles from saliva retrieved from an online genomic repository and assayed on the Illumina Human Methylation450 BeadChip microarray. These ARCpG markers were used to build a saliva-specific age prediction model that was tested in silico on an independent saliva testing data set. They were shown to perform well in terms of age prediction, and consequently, they were further validated by targeted bisulfite sequencing of additional saliva samples, using the Illumina MiSeq® platform. Subsequently, a large cohort of 754 DNAm profiles from blood samples assayed on the newly launched Illumina Methylation EPIC®BeadChip were downloaded from an online genomic repository, in order to be tested for the first time for age association. Novel AR CpG sites were identified from both the newly added probes on this chip and from probes that were found on older platforms, however, the prediction accuracy of the blood-specific age prediction model did not improve compared to models built from the older Illumina Human Methylation platforms. Finally, a multi-tissue age prediction model that is able to predict the age across the tissues was constructed. This multi-tissue age prediction model will have potential applications in forensic science in assisting investigations to predict chronological age across different biological samples,regardless of the tissue(s) those samples are derived from.
Advisor / supervisor
  • Gibson, Lorraine
  • Haddrill, Penelope R.
Resource Type
DOI
Date Created
  • 2019
Former identifier
  • 9912789393502996
Embargo Note
  • This thesis is restricted to Strathclyde users only until 1st April 2025.

关系

项目