An Approach for Analyzing Linkage Data that Accounts for Variable Levels of Heterogeneity

• Activity Number: 335
• Type: Contributed
• Date/Time: Wednesday, August 6, 2003 : 8:30 AM to 10:20 AM
• Sponsor: Biometrics Section
• Author(s): Swati Biswas*+ and Shili Lin
• Companies: The Ohio State University and The Ohio State University
• Keywords: Admixture Test ; linkage ; locus heterogeneity ; LOD score ; MCMC ; posterior probability of ;inkage

Abstract

Admixture test (A-test) is widely used for analyzing linkage data when locus heterogeneity is suspected. It is based on mixture likelihood where the mixing (heterogeneity) parameter represents the proportion of linked families. But, in general, the heterogeneity parameter varies across different types of families. Hence, the A-test estimates may be biased. We consider a new approach in which each family has its own heterogeneity parameter representing the probability that it is linked. These parameters are nuisance parameters while the main parameter of interest is the location of the disease gene. We model the problem in the Bayesian framework and use Markov chain Monte Carlo methods to estimate the posterior distributions of the parameters. The posterior probability of linkage on a chromosome is estimated. If linkage is inferred, the location of the disease gene is estimated. In that case, we also find standard error of the estimate using the batch means method. We perform diagnostics to monitor the convergence of the chain and test independence of the batches. Simulations are done to evaluate this approach under various settings including different number of disease genes.