Activity Number:
|
393
- Genetic Data Analysis, What Could Possibly Go Wrong?
|
Type:
|
Invited
|
Date/Time:
|
Wednesday, August 5, 2020 : 1:00 PM to 2:50 PM
|
Sponsor:
|
Section on Statistics in Genomics and Genetics
|
Abstract #309475
|
|
Title:
|
Too Good to Be True: Statistical Follies in Genome Wide Association Studies
|
Author(s):
|
Sharon Lutz* and Joanne Sordillo and Ann Wu
|
Companies:
|
Harvard Medical School and Harvard Medical School and Harvard Medical School
|
Keywords:
|
Genome-wide association studies;
population stratification;
collinearity
|
Abstract:
|
Genome-wide association (GWA) analyses are a useful tool for determining single nucleotide polymorphisms (SNPs) associated with complex diseases and traits. However, GWA are prone to errors due to violations of the statistical assumptions for the underlying regression models and confounding due to population substructure. While strong associations between a SNP and the trait of interest (p-value< 5e-8) are usually due to a true association, occasionally these strong associations are a spurious result due to collinearity, heteroscedasticity, or population stratification. We illustrate how to identify these issues and the corresponding solutions in the Age-Dependent Pharmacogenomics of Asthma Treatment (ADAPT) project which includes the Childhood Asthma Management Program (CAMP) study and the Genetic Epidemiology of Asthma in Costa Rica Study (GACRS) to determine SNPs associated with bronchodilator response (BDR) and exacerbations in individuals with asthma.
|
Authors who are presenting talks have a * after their name.