U.S. flag

An official website of the United States government, Department of Justice.

NCJRS Virtual Library

The Virtual Library houses over 235,000 criminal justice resources, including all known OJP works.
Click here to search the NCJRS Virtual Library

Application of six IrisPlex SNPs and comparison of two eye color prediction systems in diverse Eurasia populations

NCJ Number
International Journal of Legal Medicine Volume: 128 Dated: 2014 Pages: 447-453
Libing Yun; Yan Gu; Haseena Rajeevan; Kenneth K. Kidd
Date Published
7 pages

In this paper, the authors report on a study comparing two prediction models using the data for the six IrisPlex system, which consists of six single nucleotide polymorphisms, genotyped on 905 individuals from 12 different Eurasian populations.


DNA-based prediction for externally visible characteristics such as eye color is already a useful tool in forensic criminal investigations. The IrisPlex system, consisting of six single nucleotide polymorphisms (SNPs) and a prediction model, was developed based on individuals from several European populations. Other recent studies have developed a different prediction model, also based on European populations. In this study, the authors compared two prediction models using the data for the six IrisPlex SNPs genotyped on 905 individuals from 12 different Eurasian populations. All SNPs showed significant differences in allele frequencies among three groups of populations: European, genetically intermediate (Khanty, Uygur, and Yakut), and East Asian. The two prediction models, the FROG-kb calculation based on the formula of Walsh et al. (2011) and the Snipper calculation from Ruiz et al. (2013), gave identical predictions of brown eye color for the four East Asian populations with complete data but did not give concordant predictions for many individuals in the seven intermediate and European populations. Inconsistencies were mainly conclusive prediction by one model but not the other. Of the 714 individuals with complete 6-locus genotypes, the two models gave 22 percent inconsistent predictions. Eliminating the 306 individuals in the Korean and three Chinese populations, in which the predictions were always consistent for brown eye color, the inconsistencies (among the remaining 408 individuals) were 38.7 percent. The authors conclude that more attention should be paid to predictive uncertainty/error. Implementation of both prediction models in future forensic casework is one immediate way to highlight uncertainty. (Publisher Abstract Provided)