Whodunit? Using DNA Methylation in Forensics
For some reason crime scene investigations and murder mysteries seem to fascinate us. So much so, that the countless books and television shows never seem to get old. Like most things on television however, life in a forensics lab is often portrayed to be much more fantastical than real life may argue. No, you can’t enlarge that pixelated security photo to see the murderer’s license plate in a reflection in the window across the street from the scene of the crime. Toxicology reports take months, not hours, to get back. And no, your average forensic lab probably can’t get much information from that ONE broken strand of hair found on the ground. However, forensics technology IS progressing and recent advances may soon make the fantastical nature of these shows more of a reality.
Age prediction is extremely useful in crime investigations, whether it is used to help identify a victim or a suspect. Currently, forensic science relies on analysis of osteal markers in bones and teeth to determine age. While this method is fairly reliable, its use is restricted to cases in which a skeleton is present. Although other age biomarkers exist, such as length of leukocyte telomeres, deletions in mitochondrial DNA, and amount of advanced glycation end products, individual variations can be so large that these measurements cannot always be considered reliable.
DNA methylation is known to be correlated with age and has emerged as a promising method for age-prediction. Using DNA methylation markers in forensics however is challenging, as most methods rely on Human Methylation 450 Bead Chip technology. Although this method is accurate and reliable, it can require larger amounts of DNA, control assays, or complex bioinformatic analysis. Pyrosequencing, in contrast can use as little as 10ng of genomic DNA and has a fairly simple protocol that can be easily adapted into forensics labs.
In the current study, the authors chose to use blood as their sample of choice, since it is often found at crime scenes. Using 1415 publicly available datasets, they were able to identify both known and novel age-correlated CpG sites based on the Human Methylation 450 Bead Chip. They selected the three most informative sites, and used pyrosequencing to evaluate 765 samples from local Korean individuals. They used 535 of these samples to develop a prediction model and then validated their model on the remaining 230 samples. They found that age-correlated CpG sites were associated with genes related to development and morphogenesis. Because differentially methylated regions are known to differ slightly between male and females, they selected common sites that were not influenced by gender by comparing the top 100 age-correlated sites from both males and females and focusing only on the overlapping sites. Doing this, they settled on 25 age-related sites not influenced by gender. They did note however that the prediction accuracy of these 25 sites decreased with older individuals over 60 years.
Although the authors’ prediction model was extremely accurate in their small population, they acknowledged that they had excluded any datasets associated with disease, which may skew methylation marks. This raises the possibility that when using their method, if a blood sample found at a crime scene belongs to someone with a disease, the results may not be as accurate. Additionally, the study focused on Korean datasets and may not be as accurate in other races.
While the excitement exhibited in crime shows may still be a bit over dramatized, researchers around the world are working tirelessly to make these impressive technologies more of a reality. And soon, we will all be able to stop yelling at our TVs about how “that isn’t possible!”
Park JL, Kim JH, Seo E, Bae DH, Kim SY, Lee HC, Woo KM, & Kim YS (2016). Identification and evaluation of age-correlated DNA methylation markers for forensic use. Forensic science international. Genetics, 23, 64-70 PMID: 27017110