Group for Research in Decision Analysis

Latent Class Models for Pedigree Data

Aurelie Labbé

Considering clinical data arising from a collection of pedigrees, the general objective of our research is to develop methods to identify homogeneous disease sub-types based on multivariate disease measurements. Our approach is to develop statistical models describing the multivariate symptoms of subjects in families as a function of latent homogeneous disease classes. We extended latent class analysis to allow dependence between the latent disease class status of relatives within nuclear families. An EM algorithm maximizes the likelihood and a cross-validation approach selects the optimal model. An application of our approach to disease gene mapping by linkage analysis will be presented.