Score tests for epistasis models on quantitative traits using general pedigree data.

Efficient score tests for a two-locus model for quantitative traits are proposed for general pedigrees. Two search strategies, simultaneous search and conditional search, are considered. These tests are attractive alternatives to the likelihood ratio test for their asymptotic equivalency to their respective likelihood ratio tests and efficiency in computation. The finite-sample properties of these tests are investigated through simulations. Extensions to more complicated models and to other search strategies are discussed. Copyright 2003 Wiley-Liss, Inc.