Internal validation of the fully continuous model in EuroForMix for its implementation in routine forensic DNA...

Area of Science:

• Forensic Science
• Genetics
• Statistical Modeling

Background:

• DNA mixture analysis is crucial in forensic science but presents challenges.
• Likelihood Ratio (LR) is recommended by the International Society for Forensic Genetics (ISFG) for evaluating evidence involving persons of interest (POI).
• Mathematical models for LR calculation vary in complexity, from binary to fully continuous models incorporating peak heights.

Purpose of the Study:

• To internally validate the fully continuous model in EuroForMix (EFM) software.
• To compare EFM's performance against a semi-continuous model in LRmix Studio.
• To assess the reliability and subjectivity reduction of EFM in forensic DNA mixture analysis.

Main Methods:

• Validation using 59 artificial DNA mixtures (2-4 contributors) and 19 real forensic cases.
• Comparison of fully continuous (EFM) and semi-continuous (LRmix Studio) models.
• Analyses included contributor scrutiny, contribution estimation, hypothesis testing, sensitivity analysis, discrimination power, and software options.

Main Results:

• EFM consistently returned higher LR values than LRmix Studio in both single-source and mixed samples.
• The fully continuous model showed neutral behavior, yielding inconclusive results with scarce genetic data.
• EFM showed a higher proportion of inconclusive results when a reporting threshold was applied.

Conclusions:

• EuroForMix demonstrates correct performance, providing reliable results and reducing analytical subjectivity in forensic DNA mixture interpretation.
• The fully continuous model is a valuable tool for statistical evaluation of genetic evidence.
• Further validation and comparison with other software are essential for robust forensic applications.