Analysis of discrete phenotypes using a multipoint identity-by-descent method: application to Alzheimer's disease.

The multipoint identity-by-descent method was developed to detect linkage to a specific chromosomal region through partitioning the genetic variance. This method has previously been applied to quantitative traits, and here is extended to a qualitative trait, where a dichotomous affected/unaffected status variable is transformed to a quantitative variable by incorporating covariates. This method is applied to the Alzheimer's disease data sets from Genetic Analysis Workshop 8, to investigate putative linkage to chromosomes 19 and 21. The multipoint identity-by-descent method is used to test for linkage through the qualitative trait, and for excess sharing of the chromosomal region among affected sibs. Results are compared to those of the affected-pedigree-member method and classical linkage analysis. None of these methods gave results showing clear linkage, with the only marginally significant results occurring for the Boston data set on chromosome 19 and the Duke data set for chromosome 21 using the multipoint identity-by-descent method.