Exploratory data analysis for the interpretation of low template DNA mixtures

Area of Science:

• Forensic Genetics
• Molecular Biology
• Biostatistics

Background:

• Interpreting DNA mixtures, particularly low template DNA, is challenging due to stochastic effects like allele drop-out and drop-in.
• Classical forensic genetics approaches are insufficient for analyzing samples with such uncertainties.

Purpose of the Study:

• To develop and explore an improved probabilistic model for calculating likelihood ratios (LRs) in complex DNA mixture interpretation.
• To enhance flexibility in modeling allele drop-out and drop-in probabilities for different contributors and hypotheses.

Main Methods:

• Utilized a probabilistic model to calculate likelihood ratios (LRs) for DNA mixtures with uncertain compositions.
• Modified an existing model to allow variable drop-out probabilities per contributor and different parameters for prosecution/defense hypotheses.
• Employed Monte Carlo simulations to determine plausible ranges for drop-out probabilities.

Main Results:

• The enhanced model allows for flexible exploration of LRs by varying drop-out and drop-in parameters.
• Demonstrated the model's utility with two low template DNA mixture cases involving multiple contributors.
• The approach provides a more nuanced weight of evidence when comparing crime samples to reference samples.

Conclusions:

• The improved probabilistic model offers greater flexibility and accuracy in interpreting complex DNA mixtures, especially those with low template DNA.
• The exploratory nature of the model aids in understanding the impact of stochastic effects on the weight of evidence.
• This methodology is valuable for both high and low template DNA analysis in forensic casework.