A comparative study of qualitative and quantitative models used to interpret complex STR DNA profiles

Area of Science:

• Forensic Science
• Genetics
• Computational Biology

Background:

• Accurate interpretation of complex DNA profiles is crucial in forensic casework.
• Real-world DNA samples present challenges like mixtures, degradation, and allele sharing.
• Evaluating the performance of DNA interpretation software is essential for forensic reliability.

Purpose of the Study:

• To assess the performance of DNA profile interpretation models using realistic casework data.
• To compare the efficacy of EuroForMix, LRmix Studio, and LoCIM-tool in analyzing complex DNA mixtures.
• To evaluate likelihood ratio calculations and deconvolution capabilities of different software.

Main Methods:

• Utilized a dataset of 59 diverse, casework-realistic DNA samples, including mixtures and degraded profiles.
• Performed weight-of-evidence (likelihood ratio) analyses comparing LRmix Studio and EuroForMix.
• Conducted deconvolution analyses comparing EuroForMix and LoCIM-tool to identify unknown contributors.

Main Results:

• EuroForMix exhibited a higher true positive rate with comparable false positive rates versus LRmix Studio.
• EuroForMix showed improved discrimination in mixture cases, especially when the person of interest was a minor contributor.
• EuroForMix outperformed LoCIM-tool in the deconvolution of major contributors in DNA profiles.

Conclusions:

• EuroForMix is a more effective tool for interpreting complex DNA profiles in forensic applications.
• The software's ability to handle mixtures and identify minor contributors enhances its utility in casework.
• These findings support the use of advanced probabilistic genotyping software for robust DNA evidence interpretation.