Improving the prediction of human quantitative pigmentation traits such as eye, hair and skin color using a worldwide representation panel of US and European individuals

As submitted by the proposer: Forensic DNA Phenotyping or 'DNA intelligence' tools can significantly aid police investigations towards finding unknown individuals, by providing physical appearance information on unknown suspects, perpetrators and missing persons from biological samples. This is especially useful in cases where conventional DNA profiling (STR typing) remains non-informative and an investigation stalls. Until now, categorical eye and hair color were the only traits that could be predicted from DNA with adequate accuracy, using the HIrisPlex system as an example. However, certain categories such as green or intermediate eye color, individuals with an age-dependent hair colour change such as blond to brown/black, or admixed individuals, could not be predicted to the same levels of accuracy required for optimal physical appearance prediction from DNA. To circumvent these predictive issues, it is essential to move away from categorical (groups such as blue or brown) towards quantitative Genome Wide Association Studies (GWAS) to generate enough information for a more individualized prediction outcome from DNA. When an outcome of individualized prediction is successful, this could change how standard DNA profiling is performed on difficult (no hit on the database) casework scenarios within the United States and worldwide. The first aim of this study is to carry out new quantitative GWAS searches on 2000 individuals from three countries that have been enriched with the problem areas mentioned above and that have affected the categorical prediction of eye, hair and skin colour to date. This 2000 set of Irish, Greek and US individuals contain a well represented spectrum of variable pigmentation for all three quantitative traits eye, hair and skin colour. The second aim of this study is to generate a new phenotype-genotype database and prediction model from an additional 3000 individuals from the same three countries that is more representative of a worldwide population set than has previously been used for modeling parameters in the prediction of pigmentation traits. A collaboration that involves several populations ensures a broad representation of the common and unique alleles found associated and predictive for these quantitative traits, irrelevant of ancestry. The third aim of this study is to create a computerised prediction system that is capable of producing a photo print out of quantitative colour for eye, hair and skin colour prediction from DNA. This eliminates any interpretive issues that were previously associated with categorical pigmentation prediction and brings individualized physical appearance prediction to a new and exciting level in investigative forensic biology. ca/ncf