Evaluation of mixed-source, low-template DNA profiles in forensic science

Area of Science:

• Forensic Science
• Genetics
• Statistical Modeling

Background:

• Advances in DNA profiling allow analysis of minute DNA quantities (tens of picograms), even from degraded samples.
• Low-template DNA (LTDNA) analysis presents challenges due to stochastic effects like allele dropout, dropin, and variable stutter peaks.
• The reliability of LTDNA evidence has faced scrutiny in high-profile legal cases, impacting its admissibility in court.

Purpose of the Study:

• To propose a robust statistical method for evaluating low-template DNA (LTDNA) profile evidence.
• To address the limitations of existing statistical approaches in handling the complexities of LTDNA data.
• To provide a tool that enhances the reliability and interpretability of forensic DNA evidence.

Main Methods:

• Development of a statistical model accounting for multiple replicates with varying dropout rates and sporadic dropins.
• Incorporation of parameters for different contributor DNA amounts, relatedness between individuals, uncertain allele designations, and sample degradation.
• Implementation of the method using open-source R code for transparency and broad accessibility.

Main Results:

• The proposed statistical method demonstrates effective performance in evaluating complex LTDNA profiles.
• Illustrative examples using real forensic cases and simulated crime scene profiles validate the method's utility.
• The approach provides a more reliable framework for interpreting noisy DNA evidence.

Conclusions:

• The validity of statistical evaluation methods is crucial for the acceptance of LTDNA evidence in legal proceedings.
• The developed statistical framework offers a powerful tool to overcome the challenges posed by LTDNA analysis.
• This method can enhance the probative value of forensic DNA evidence, ensuring its appropriate use in the justice system.