Identifying common donors in DNA mixtures, with applications to database searches

Area of Science:

• Forensic Science
• Genetics
• Statistics

Background:

• Likelihood ratio (LR) calculations are crucial for evaluating DNA evidence, particularly for mixed traces involving multiple contributors.
• Existing LR methods primarily focus on comparing a single mixed trace (M) to a person of interest's genotype (g).

Purpose of the Study:

• To generalize the LR calculation to compare two mixed DNA traces (M1, M2) to determine if they share a common donor.
• To establish the method's effectiveness in identifying common contributors in forensic DNA analysis.

Main Methods:

• Developed a generalized likelihood ratio LR(M1, M2) for comparing two mixed DNA traces.
• Utilized a semi-continuous mixture evaluation model with simulated data to derive Receiver Operating Characteristic (ROC) curves.
• Demonstrated efficient database search strategies for comparing multiple mixed traces.

Main Results:

• The generalized LR method effectively detects common donors, achieving a very small false positive rate with a high detection probability.
• ROC curves indicate the feasibility of common donor searches across various mixture types.
• The method was successfully applied to mixed DNA traces from the Dutch DNA database, showcasing its practical efficiency.

Conclusions:

• The novel LR(M1, M2) method provides a robust statistical framework for identifying common donors in complex forensic DNA mixtures.
• The approach is computationally efficient, enabling large-scale database searches and improving the investigative power of DNA analysis.