A new statistic for the analysis of association between trait and polymorphic marker loci

Inference for detecting the existence of an association between a diallelic marker and a trait locus is based on the chi-squared statistic with one de gree of freedom. For polymorphic markers with m alleles (m > 2), three appr oaches are mainly used in practice. First, one may use Pearson's chi-square d statistic with m - 1 degrees of freedom (d.f.) but this leads to a loss i n test power. Second, one can select an allele to be the most associated an d then collapse the other allele categories into a single class. This reduc es in a biased way, the locus to a diallelic system. Third, one may use the Terwilliger [J.D. Terwilliger, Am. J, Hum. Genet. 56 (1995) 777] likelihoo d ratio statistic which has a non-standard unknown limiting probability dis tribution. In this paper, we propose a new statistic, L-D, based on the sec ond testing approach. We derive the asymptotic probability distribution of L-D in an easy way. Simulation studies show that L-D is more powerful than Pearson's chi-squared statistic with m - 1 d.f. (C) 2000 Elsevier Science I nc. All rights reserved.