Mixture detection with Demixtify.

Area of Science:

• Forensic Science
• Genetics
• Bioinformatics

Background:

• Short tandem repeats (STRs) are standard forensic genetic markers, with established tools for mixture analysis.
• Forensic genetic genealogy increasingly uses single nucleotide polymorphisms (SNPs), but analytical tools for SNP mixture assessment are limited.

Purpose of the Study:

• Introduce Demixtify, a novel computational tool for detecting DNA mixtures using biallelic SNPs.
• Address the lack of specialized tools for forensic SNP mixture analysis.

Main Methods:

• Developed Demixtify, a program designed to analyze biallelic SNPs for DNA mixture detection.
• Incorporated an empirical estimator of sequence error specific to the assayed markers.
• Developed orthogonal techniques to characterize in vitro mixtures and single-source samples for validation.

Main Results:

• Demixtify successfully detects highly imbalanced mixtures (down to 1:99) with ample coverage.
• The program can identify mixtures in low-coverage (∼1×) samples, particularly when balanced.
• Orthogonal methods validated the effectiveness of Demixtify's mixture detection capabilities.

Conclusions:

• Demixtify offers a powerful solution for DNA mixture detection using SNPs in forensic applications.
• The tool enhances the capabilities of forensic genetic genealogy by addressing SNP mixture analysis challenges.
• The inclusion of a marker-specific error estimator increases its relevance and reliability for forensic casework.