Relationship inference based on DNA mixtures.

Area of Science:

• Forensic Genetics
• Statistical Genetics
• Computational Biology

Background:

• Kinship analysis typically relies on established DNA profiling tools.
• Challenges arise in kinship cases involving DNA mixtures and absent reference profiles, such as in forensic investigations.
• Existing methods are often inconvenient or inapplicable when reference data is missing.

Purpose of the Study:

• To develop statistical methods for general relationship inference using DNA mixtures.
• To address kinship cases where reference profiles are unavailable.
• To provide a practical tool for analyzing complex DNA mixtures in forensic contexts.

Main Methods:

• Decomposition of likelihood calculations for mixtures into a series of kinship problems.
• Development of statistical methods to handle mutations, silent alleles, and θ-correction.
• Implementation of these methods in the freely available R package relMix, extending Familias.

Main Results:

• The relMix package effectively handles general relationship inference from DNA mixtures.
• Statistical methods accommodate complicating factors like mutations and silent alleles.
• The approach is demonstrated using a real-world case involving a rape victim and fetal DNA.

Conclusions:

• The presented statistical methods and relMix package offer a robust solution for kinship inference from DNA mixtures, even with missing reference data.
• This work facilitates the analysis of complex forensic DNA samples.
• The developed tools enhance the utility of kinship analysis in challenging forensic scenarios.