Genetic analysis of von Willebrand's disease in two large pedigrees: a multivariate approach

Clinical and laboratory data, including polymorphic marker traits for linkage analysis, were collected from two large multigenerational families segregating for von Willebrand disease. A new approach to the identification of gene carriers in these families, combining pedigree segregation analysis with multivariate discriminant analysis, is applied. Whereas individually the clinical symptoms and the factor VIII related activities could not distinguish between hypotheses, it was possible to find a discriminant function-showing consistency of the data with a dominant gene hypothesis, but not with a recessive gene or an environmental hypothesis. This function is estimated to lead to 3.2% and 5.5% minimum misclassification of the genotypes, respectively, in the two families. The discriminant function could be used for other families, but is should be calibrated for the specific population in which it is used. Among the markers investigated, GPT is the most likely to be linked to von Willbrand's disease, with a maximum lod score of about unity at 15% recombination.