A new robust AI/ML based model for accurate forensic age estimation using DNA methylation markers.

Area of Science:

• Genetics and Forensic Science
• Epigenetics and Molecular Biology

Background:

• Epigenetic markers, specifically DNA methylation at CpG sites, are crucial for accurate age prediction.
• These markers are increasingly vital in forensic science for criminal investigations and human identification.

Purpose of the Study:

• To identify a minimal set of CpG sites for robust and accurate forensic age estimation.
• To validate the predictive accuracy of selected CpG sites across diverse populations and conditions.

Main Methods:

• Screened 476,366 CpG sites based on high correlation with chronological age (>0.85) and low mutual correlation (<0.15).
• Selected 12 CpG sites associated with genes including FHL2, ELOVL2, and TRIM59.
• Utilized three different neural network models to assess predictive accuracy, independent of model specifics.

Main Results:

• Identified 12 CpG sites demonstrating high accuracy for age prediction.
• Neural network models achieved comparable accuracy across diverse populations, genders, and health statuses.
• Validation on independent datasets confirmed the reliability of the selected CpG sites for age estimation.

Conclusions:

• A minimal set of 12 CpG sites provides a robust and efficient solution for forensic age estimation.
• This approach significantly enhances the precision and reliability of forensic investigations.
• The findings support the use of epigenetic markers for accurate age prediction in forensic contexts.