Multipoint Linkage Disequilibrium Mapping Using Multilocus Allele Frequency Data

Abstract

This paper describes a likelihood based fine scale linkage disequilibrium mapping method for estimating the position of a disease predisposing gene relative to a battery of typed marker loci. The method uses multilocus allele frequency data from a sample of unrelated diseased individuals and from a sample of unrelated control individuals, that is, a case and control type design. This type of data could be obtained by typing DNA pools, which is less expensive than typing individuals separately. The method described uses a nonparametric model that makes it robust to the shape of the genealogy at the disease locus. It can be implemented efficiently, making a multipoint analysis of a data set of a thousand markers feasible. An example power analysis uses simulations to estimate the amount of information that can be extracted from fully resolved haplotype data, relative to multilocus allele frequency data. For the assumed parameter values and a battery of 10 markers, roughly three times narrower region estimates can be derived from haplotype data than from allele frequency data only. Depending on how we choose to measure information, allele frequency data at an additional approximately 18 or approximately 33 markers is needed to compensate for this loss of information.