A new implementation of a semi-continuous method for DNA mixture interpretation

Area of Science:

• Forensic Science
• Genetics
• Computational Biology

Background:

• Accurate Likelihood Ratio (LR) calculation is crucial for interpreting DNA mixture profiles in forensic science.
• Existing methods may have limitations in handling population structure, allelic drop-out, and drop-in.
• The CODIS software package's PopStats module is a key tool for DNA mixture analysis.

Purpose of the Study:

• To introduce and validate a new calculation module, SC Mixture, for assigning the Likelihood Ratio (LR) of DNA mixture profiles.
• To implement a semi-continuous model that incorporates population structure, allelic drop-out, and drop-in.
• To evaluate the performance and consistency of the SC Mixture module against established software (PopStats, MixKin, LRmix).

Main Methods:

• Developed a new semi-continuous model (SC Mixture) within the PopStats module, based on MixKin software mathematics.
• The model integrates over possible drop-out rates for each contributor, specifying user-defined drop-in rates and population structure (theta).
• Conducted an extensive validation study on ten mixtures with 1-5 contributors, simulating drop-out and evaluating 1620 combinations across different parameters.

Main Results:

• The SC Mixture module successfully assigned Likelihood Ratios (LR) for DNA mixtures with varying numbers of contributors and simulated drop-out.
• The new module demonstrated considerable consistency in results when compared with PopStats, MixKin, and LRmix across all tested scenarios.
• The semi-continuous model effectively handles population structure and allelic drop-out/drop-in without requiring laboratory-specific parameters like peak heights.

Conclusions:

• The SC Mixture module provides a robust and validated method for calculating Likelihood Ratios in complex DNA mixtures.
• This new implementation offers a valuable tool for forensic DNA analysis, enhancing the interpretation of mixture profiles.
• The consistency across different software packages supports the reliability of the semi-continuous modeling approach for DNA mixture interpretation.