4:30 PM-5:15 PM
Chair: Leon Glass, McGill University, Canada
Room: Convocation Hall
In recent years, much progress has been made in the localization of disease genes on the human gene map, which represents the first step in the so-called positional cloning approach to elucidating the etiology of genetic traits. Disease gene mapping rests on observations of family members, i.e. nonindependent data. Parameters such as recombination fractions and map positions are generally estimated by maximum likelihood. A major technical problem is the calculation of the likelihood, which is always done in a numerical fashion, and MLEs are obtained iteratively. Also, for many diseases, simplified inheritance models must be assumed in order for likelihood calculations to be feasible. To avoid these modeling problems, nonparametric approaches have been developed. The speaker will present various approaches to gene mapping, and will discuss problems yet to be resolved.
Laboratory of Statistical Genetics, Rockefeller University and Department of Genetics and Development, Columbia University