Application of an ordered subset analysis approach to the genetics of alcoholism.

For complex diseases, underlying etiologic heterogeneity may reduce power to detect linkage. Thus, methods to identify more homogeneous subgroups within a given sample in a linkage study may improve detection of putative susceptibility loci. In this study we describe an ordered subsetting approach that utilizes disease-related quantitative trait data to complement traditional linkage analysis. This approach uses family-based lod scores derived from the initial genome screen and a family-based descriptor of the trait of interest. The goal of the approach is to identify more homogeneous subgroups of the data by ranking families based on their quantitative trait data. Permutation testing is used to assess statistical significance. This approach can be adapted to a variety of linkage methods and may provide a means to dissect some of the underlying heterogeneity in complex disease genetics.