Evaluation of the IrisPlex system for eye colour prediction in the Serbian population

Abstract

DNA-based prediction of a physical appearance, also known as DNA phenotyping, is a rapidly developing field with great potential for solving difficult forensic investigations. Externally visible characteristics (EVCs), such as eye colour, are easily recognised and genetically determined. Analysis of highly informative single nucleotide polymorphisms (SNPs) encompassing trait-associated genes provides information about the phenotype of an unknown individual, which is critical in cases where the standard STR profile is not useful. The IrisPlex assay was developed based on genotype data from 3804 Dutch Europeans, with the goal of accurately predicting brown, blue, and undefined eye colours using the six SNP markers alone. This assay has been validated in several studies and has shown high accuracy of prediction for brown and blue eye colours in most European populations. The aim of this work was to evaluate IrisPlex in the Serbian population and to determine if there are possible discrepancies in prediction accuracy compared to the previously published European data. Therefore, we performed the IrisPlex analysis on 65 Serbian individuals. Our results showed that the prediction of eye colour was accurate for about 57% of the respondents, which did not change significantly by applying the 0.7 threshold. The IrisPlex system performed well in predicting blue and brown eye colour in the Serbian sample, achieving high sensitivity levels. However, this model appeared to be unsuitable in the prediction of undefined eye colour, which was a common phenotype in our sample. This low sensitivity suggests a diverse and possibly population-specific genetic background of undefined eye colour. Given the large genetic diversity of the Serbian population, further work on a larger Serbian sample with more SNPs analysed is needed to reveal genetic variants mainly associated with this trait.