Statistical methods for population genetics are finding increasing areas of application, including DNA profiles, evolutionary phylogenies, hereditary disease patterns based on mitochondrial DNA, and even the origins and spread of language.

DNA profiles of several genetic loci are an important tool for identification to measure alleles in hypervariable loci, whether based on RFLP or PCR techniques. The use of these techniques, however, to ascertain similarity of two profiles and to estimate the relative frequency of a profile in a specified population and their underlying statistics have been the subject of much heated debate.

The building of evolutionary trees to establish phylogenies, or the relations between organisms, families, or species, is another important area of population genetics. In the medical sciences this is particularly important in the area of microbiology and parasitology. Examples include malaria, HIV, and tuberculosis. It is also used in studies of descent based on mitochondrial DNA. Mitochondrial DNA is inherited from the mother, so requires different statistical approaches than the usual linkage analysis based on nuclear DNA. Several difficult statistical issues arise in phylogeny building. One is the alignment of sequences. Second is the estimation of the distances based on some assumed evolutionary model. Third is the building of a tree structure based on a clustering algorithm. Fourth, is the problem of inference about the complex tree structure that was obtained. Methods of inference include likelihood methods and bootstrapping, but neither is well-understood in this application.

The statistical methods of population genetics and for the analysis of disease patterns based on mitochondrial DNA have lagged behind those for linkage analysis based on nuclear DNA. This workshop will be a unique opportunity for top-level statisticians to meet with quantitative geneticists to advance the statistical underpinnings of these emerging problems.

top of page