NCJ Number
246370
Journal
Forensic Science International: Genetics Volume: 7 Issue: 2 Dated: February 2013 Pages: 251-263
Date Published
February 2013
Length
13 pages
Annotation
The aim of this paper is to establish a LR-based framework using principles of the basic model recommended by the ISFG DNA commission.
Abstract
Although likelihood ratio (LR) based methods to analyze complex mixtures of two or more individuals, that exhibit the twin phenomena of drop-out and drop-in has been in the public domain for more than a decade, progress towards widespread implementation in to casework has been slow. Researchers use the tools in the form of open-source software (LRmix) in the Forensim package for the R software. A generalized set of guidelines has been prepared that can be used to evaluate any complex mixture. In addition, a validation framework has been proposed in order to evaluate LRs that are generated on a case-specific basis. This process is facilitated by replacing the reference profile of interest (typically the suspect's profile) with simulated random man using Monte-Carlo simulations and comparing the resulting distributions with the estimated LR. Validation is best carried out by comparison with a standard. Because LRmix is open-source the researchers proposed that it is ideally positioned to be adopted as a standard basic model for complex DNA profile tests. This should not be confused with 'the best model' since it is clear that improvements could be made over time. Nevertheless, it is highly desirable to have a methodology in place that can show whether an improvement has been achieved should additional parameters, such as allele peak heights, are incorporated into the model. To facilitate comparative studies, the researchers provide all of the necessary data for three test examples, presented as standard tests that can be utilized to carry out comparative studies. The researchers envisage that the resource of standard test examples will be expanded over coming years so that a range of different case-types that are included will be used in order to improve the efficacy of models; to understand their advantages; conversely, to understand any limitations and to provide training material. (Published Abstract)