|
Activity Number:
|
320
|
|
Type:
|
Topic Contributed
|
|
Date/Time:
|
Tuesday, August 8, 2006 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Biometrics Section
|
| Abstract - #307294 |
|
Title:
|
Latent Variable Mixture Modeling with Genetic Applications
|
|
Author(s):
|
Bengt Muthen*+
|
|
Companies:
|
University of California, Los Angeles
|
|
Address:
|
Box 951521, Los Angeles, CA, 90095,
|
|
Keywords:
|
latent class hybrid analysis ; QTL ; IBD ; twins ; siblings
|
|
Abstract:
|
Modeling of data from twins, siblings, and other family members rely on a well-measured phenotype. In many applications, the phenotype is best characterized as a latent variable measured by a set of observed, fallible indicators. This presentation will propose a factor mixture (finite mixture) model with latent classes representing different response profiles corresponding to different subtypes of disorders and continuous variables representing severity variation within type. Genetic influence on both class membership and severity will be considered. Maximum-likelihood estimation with categorical outcomes using EM and numerical integration will be discussed in the Mplus framework. Applications include sibling data concerning ADHD problems. The sibling data analysis uses quantitative trait locus analysis with genetic correlations varying by IBD proportion.
|
- The address information is for the authors that have a + after their name.
- Authors who are presenting talks have a * after their name.
Back to the full JSM 2006 program |
