DNA methylation-based age prediction with bloodstains using pyrosequencing and random forest regression

Area of Science:

• Forensic science
• Epigenetics
• Genomics

Background:

• DNA methylation is a promising biomarker for predicting chronological age in forensic investigations.
• Existing age prediction models often require high DNA quantities, limiting their forensic applicability, especially with trace evidence like bloodstains.

Purpose of the Study:

• To develop a sensitive and rapid DNA methylation-based assay for age prediction in bloodstains suitable for forensic casework.
• To evaluate the efficacy of pyrosequencing and random forest regression for age estimation from limited DNA samples.

Main Methods:

• Developed a pyrosequencing assay to analyze methylation levels at 46 CpG sites across six genes in bloodstain samples.
• Utilized random forest regression to build separate age prediction models for males and females using data from 128 males and 113 females (aged 10-79).
• Optimized models using seven key CpG sites (3 for males, 4 for females) achieving high accuracy.

Main Results:

• The assay requires only 0.1 ng of genomic DNA and can be completed within 10 hours.
• The developed age prediction models achieved a mean absolute deviation of less than 3 years for both males and females.
• Identified specific CpG sites crucial for accurate age prediction in bloodstains.

Conclusions:

• DNA methylation analysis using pyrosequencing and random forest regression provides a practical and accurate method for forensic age estimation from bloodstains.
• This technique enhances forensic investigations by enabling reliable biological age determination from limited and degraded DNA samples.
• The developed assay offers a valuable tool for forensic science, improving the information obtainable from trace evidence.