Kinship cases with partially specified hypotheses.

Area of Science:

• Forensic genetics
• Statistical genetics
• Bioinformatics

Background:

• Forensic kinship testing statistically compares genetic data to hypothesize relationships.
• Standard likelihood ratio (LR) calculations can be challenging with uncertain relationships or multiple possible pedigrees.
• Complex scenarios, such as disaster victim identification (DVI), often involve incomplete familial information.

Purpose of the Study:

• To introduce a generalized likelihood ratio (GLR) to address limitations in standard LR calculations for forensic kinship testing.
• To explore the application and properties of the GLR in complex kinship scenarios, including DVI and pedigree data validation.

Main Methods:

• Definition of the generalized likelihood ratio (GLR) as the ratio of maximal likelihoods for competing hypotheses.
• Application of the GLR to discrete alternatives in forensic kinship testing and DVI.
• Utilizing the GLR over a continuous parameter space for pedigree data correctness checks.

Main Results:

• The GLR provides a robust method for statistical comparison in kinship testing, even with ambiguous pedigree information.
• Demonstrated utility of the GLR in resolving and reporting results for complex disaster victim identification cases.
• The GLR serves as an effective quality control tool for verifying pedigree data accuracy in genetic studies.

Conclusions:

• The generalized likelihood ratio (GLR) offers a flexible and powerful statistical framework for forensic kinship analysis.
• GLR enhances the ability to handle complex familial relationships and uncertain data in forensic investigations.
• This statistical approach is valuable for both direct kinship determination and ensuring data integrity in genetic research.