EFMrep

EFMrep extends the EuroForMix probabilistic genotyping model to combine and analyze Short Tandem Repeat (STR) DNA mixture samples from different multiplexes for forensic likelihood ratio calculation and deconvolution of contributors.

Key Features:

• Extended Model Parameters: Assigns distinct model parameters to each individual DNA profile within an analysis.
• Combination of STR Samples: Combines STR DNA mixture samples originating from multiple multiplexes into a single joint analysis.
• Likelihood Ratios and Deconvolution: Calculates combined likelihood ratios and performs deconvolution to separate individual contributor profiles within mixed samples.
• Incorporation of Related Unknowns: Allows specification of related unknown individuals to incorporate kinship propositions into mixture interpretation.
• Empirical Validation: Validated using the PROVEDIt dataset, demonstrating information gain when multiple samples are combined across multiplexes.
• Comparative Advantage: Generalizes the EuroForMix modeling approach to support per-profile parameterization and multi-multiplex integration.

Scientific Applications:

• Forensic DNA mixture interpretation: Interpret complex STR mixtures involving multiple contributors by computing likelihood ratios and deconvolution results.
• Multi-multiplex sample integration: Integrate samples from different multiplexes to increase information content and improve inference across datasets.
• Kinship and relatedness-aware analyses: Test propositions involving related unknowns and incorporate kinship structures into mixture evaluation.

Methodology:

Assigns per-profile model parameters, combines STR mixture data across multiplexes, computes combined likelihood ratios, performs deconvolution of contributors, and models related unknowns, with empirical validation on the PROVEDIt dataset.