A NEW TEST STATISTIC FOR LINKAGE APPLIED TO BIPOLAR DISORDER AND MARKER D18S41

We applied a new test statistic for linkage that removes the tradition al assumption of equal female (theta(f)) and male (theta(m)) recombina tion fractions by testing H-o:theta(f)+theta(m) = 1 vs. H-A:theta(f)+t heta(m) < 1 to GAW10 Problem 1. Specifically we reanalyzed the reporte d possible linkage between a suggested susceptibility locus for bipola r affective disorder and marker D18S41 on chromosome 18 [Stine et al., 1995]. We used penetrance functions fitting the description of those used by Stine et al. [1995] assuming a continuous age-dependent logist ic distribution. Maximum likelihood marker allele frequencies were est imated assuming Hardy-Weinberg equilibrium. Results from the tradition al led-score analyses do not strongly support the existence of linkage between the disease locus and marker D18S41. Similarly, the new test statistic for linkage failed to provide evidence in support of linkage . This was true whether dominant or recessive models of inheritance we re assumed, and whether the analyses included all available pedigrees or were confined to paternally transmitted pedigrees. The appreciable difference found between our led scores and those obtained by Stine et al. [1995] can be attributed to differences in the assumptions made r egarding the age-dependent penetrance function, the marker allele freq uencies, or both. (C) 1997 Wiley-Liss, Inc.