Related Experiment Videos
Comparison of multivariate tests for genetic linkage
1Departments of Epidemiology and Biomathematics, The University of Texas M.D. Anderson Cancer Center, Houston, Tex., USA. camos@request.mdacc.tmc.edu
Human Heredity
|February 15, 2001
Summary
Bivariate linkage analysis using variance component (VC) methods is more powerful than univariate tests for complex traits. The unconstrained bivariate VC test is recommended for preliminary linkage analysis with multivariate phenotypes.
Area of Science:
- Genetics
- Statistical Genetics
- Bioinformatics
Background:
- Multivariate linkage analysis offers increased power compared to univariate methods.
- Previous studies have not comprehensively compared type I error rates and power of common multivariate linkage methods.
Purpose of the Study:
- To compare the performance of bivariate formulations of variance component (VC) and Haseman-Elston (H-E) approaches.
- To evaluate type I error rates and comparative power of these bivariate methods against univariate approaches.
Main Methods:
- Simulation studies were conducted to compare bivariate H-E tests with unconstrained and constrained bivariate VC approaches.
- Performance was assessed against univariate methods.
Main Results:
- Bivariate methods demonstrated superior power over univariate analyses, except when traits exhibited very high positive correlation.
- The bivariate H-E test showed lower power than VC procedures.
- Constrained VC tests were generally less powerful than unconstrained versions.
- Empirical distributions of bivariate H-E and unconstrained bivariate VC tests aligned with asymptotic distributions for sample sizes of 100+ sibships (size 4).
Conclusions:
- The unconstrained VC test is a valuable tool for preliminary linkage testing with multivariate phenotypes.
- Bivariate VC tests generally outperformed the bivariate H-E test in terms of power.