|
Activity Number:
|
64
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Sunday, August 6, 2006 : 4:00 PM to 5:50 PM
|
|
Sponsor:
|
Section on Statistics in Epidemiology
|
| Abstract - #306494 |
|
Title:
|
Correcting for Measurement Errors in Structured Association Tests
|
|
Author(s):
|
Jasmin Divers*+ and Laura K. Vaughan and David Redden and Jose R. Fernandez and David B. Allison
|
|
Companies:
|
The University of Alabama at Birmingham and The University of Alabama at Birmingham and The University of Alabama at Birmingham and The University of Alabama at Birmingham and The University of Alabama at Birmingham
|
|
Address:
|
1665 University Blvd., Birmingham, AL, 35294,
|
|
Keywords:
|
structure association tests ; linkage desequilibrium ; measurement error ; population stratification ; genetic background variables ; ancestry estimation
|
|
Abstract:
|
Although long blocks of linkage disequilibrium in admixed individuals offer the potential for admixture mapping, admixture can confound association testing. Confounding occurs when the allele frequencies of markers to be tested and the distribution of the phenotype of interest differ among the parental populations. Various methods have been proposed to control for such genetic heterogeneity by estimating a measure of genetic background (GB) to be used as covariate in the test of association. However, these approaches do not account for variation between estimated and true GB values. Including variables measured with error in a regression analysis reduce the ability to control for confounders resulting in a higher than expected type I error rate. We present several measurement error correction methods applicable to this problem and conduct simulation studies to compare their efficacy.
|
