Prediction of eye color from genetic data using Bayesian approach

Area of Science:

• Forensic genetics
• Human pigmentation genetics
• Computational biology

Background:

• Visible traits, such as eye color, can be predicted from genetic data, potentially accelerating forensic investigations.
• Genetic research has identified numerous polymorphisms linked to human iris color variation.
• Single nucleotide polymorphisms (SNPs) are key genetic markers for observable traits.

Purpose of the Study:

• To develop and evaluate Bayesian network models for predicting eye color from genetic data.
• To assess the accuracy of different eye color classification schemes (four-category vs. two-category).
• To determine the efficacy of using specific genetic predictors for forensic applications.

Main Methods:

• Development of two Bayesian network models utilizing genetic data for eye color prediction.
• Input genetic data comprised six single nucleotide polymorphisms (SNPs) known to influence eye color.
• Evaluation of model accuracy using distinct eye color categorizations: blue, brown, green, hazel versus light and dark.

Main Results:

• High accuracy was achieved in predicting eye color, particularly with the simplified two-state classification (light vs. dark).
• The study demonstrated that reliable differentiation between light and dark irises is feasible using the selected genetic markers.
• Bayesian network analysis proved effective for interpreting genetic evidence related to pigmentation.

Conclusions:

• Genetic prediction of eye color, specifically differentiating light from dark irises, is highly accurate using six SNPs and Bayesian networks.
• This approach offers a valuable tool for forensic investigations requiring rapid identification of traits.
• Further research can refine these models for broader forensic applications in trait prediction.