|
Activity Number:
|
426
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Wednesday, August 9, 2006 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Biometrics Section
|
| Abstract - #306807 |
|
Title:
|
A Mixture Model Approach in Analyzing Genotype-Phenotype Association
|
|
Author(s):
|
Jason Robarge*+ and Lang Li and David Flockhart
|
|
Companies:
|
Indiana University School of Medicine and Indiana University and Indiana University School of Medicine
|
|
Address:
|
1050 Wishard Blvd., RG4101, Division of Biostatistics, Indianapolis, IN, 46202,
|
|
Keywords:
|
mixture model ; pharmacogenetics ; genetic association
|
|
Abstract:
|
The motivation for pharmacogenetic studies is the potential to guide individual-specific therapy based on a clinical phenotype predicted from genetic variation. To detect true genotype-phenotype associations, analysis methods must control inflation of family wise type-I error in multiple comparisons and effectively reduce data dimensionality, while maintaining statistical power. A family of mixture models has been developed to analyze genotype-phenotype associations in genetic association studies. This method can classify a continuous or binary phenotype based on genotype and reduce the multiple-comparison burden when many genotypes are compared. The mixture model is an innovative way to formulate a multiple-comparison problem into a cluster analysis, which possesses real advantages over current methods.
|
