PRS
Polygenic risk scores (PRS) are statistical models that estimate an individual's propensity for trait of interest from his or her genotype.
Construction of PRS
As is standard in statistical learning, PRS are constructed via a fitting procedure. We use training data consisting of the genotypes and trait values of individuals from a population of interest to build a model relating genotype to trait value. This model than then be used to predict trait values from genotypes for previously-unseen individuals.
Applications of PRS
The core clinical application of PRS is in risk stratification. For example, an individual with a very high PRS score for heart disease might be given a preventive treatment regime even when completely asymptomatic.