EFMex: Using EuroForMix to evaluate DNA mixtures with multiple persons of interest

We present EFMex (EuroForMix-Exhaustive), an R package for calculating likelihood ratios (LRs) in DNA mixture interpretation involving multiple persons of interest (POIs). The software implements an exhaustive method framework, evaluating all possible subsets of POIs as contributors to the mixture. We demonstrate the importance of this framework, particularly when contributors, such as close relatives, exhibit high allele sharing. A series of simulation experiments, designed to reflect complex casework scenarios, were conducted using three- and four-person mixtures in which at least two contributors were in the same family. The exhaustive method proved effective in distinguishing true contributors from closely related non-contributors, even under challenging conditions. A recalculation of exhaustive LR for candidates obtaining LR>1 with the exhaustive method was used to increase discrimination further. An alternative to the exhaustive method, using a generalized likelihood ratio, returned values very close to those of the exhaustive method. A Shiny app was created as a graphical user interface to ease the analysis for practitioners.