A fairly general construction of the polygenic score has been described, including weighted and unweighted methods from single marker analysis, and shrinkage methods used in multivariate analysis. There is little to choose between these estimators in terms of power, correlation or AUC, but the unweighted estimator will perform relatively worse as sample size increases since its sampling error does not reduce to zero. Shrinkage estimation leads to reduced mean square error for prediction and has some other advantages [35], [36], but in the main applications for polygenic scores to date, namely association testing and AUC, it does not improve over the linear regression estimate.
