Title: "Dissecting the ApoE locus with respect to Alzheimer's disease using novel association statistics"
Speaker: Dr. Stefan Bohringer, Institut fr Humangenetik, Universitt Duisburg-Essen; Essen, Germany
Place: Stanely Coulter (SC) 239; Tuesday, 4:30pm

Abstract

In complex disorders fine mapping remains a major challange in regions of high linkage disequilibrium (LD). Methods to characterize LD in high risk haplotypes can help to define causative variants. We study a likelihood approach to estimate parameters characterizing disease predisposition in a given region, i.e. haplotype frequencies, penetrance. The likelihood frameworks allows to construct tests and confidence intervals for the parameters involved. Our model allows for constistent simultaneous estimation of LD and penetrance if an unconditional sampling scheme is assumed. We illustrate the methods on a data set of Alzheimer's for which a causative locus has been demonstrated (Martin et al. 2000). It is shown that the effect of a local causative variant can be decomposed from that of unobserved, unlinked loci. To mitigate the assumptions made in the likelihood approach we show nonparametric alternatives based on contingency tables and asymptotic normal theory. These tests are less powerfull than the likelihood ratio tests but are robust against violations of certain assumptions made in the likelihood setting.



See http://www.stat.purdue.edu/~doerge/BIOINFORM.D/FALL04/sem.html for a full scheule of BIOINFORMATICS SEMINARS.