Related Experiment Videos
Genetic association analysis under complex survey sampling: the Hispanic Community Health Study/Study of Latinos
Dan-Yu Lin1, Ran Tao1, William D Kalsbeek1
1Department of Biostatistics, University of North Carolina, Chapel Hill, NC 27599-7420, USA.
American Journal of Human Genetics
|December 7, 2014
Summary
This study introduces weighted estimators to accurately analyze genetic associations in complex cohort studies. These methods correct for unequal sampling and participant relatedness, improving genetic discovery in diverse populations.
Area of Science:
- Genetics
- Epidemiology
- Statistical genetics
Background:
- Cohort studies are valuable for exploring genetic influences on diseases and traits.
- Complex sampling designs, common in cohort studies, can introduce bias if not accounted for.
- Ignoring participant relatedness within clusters can lead to inaccurate genetic association estimates.
Purpose of the Study:
- To develop and validate weighted statistical estimators for genetic association studies in complex cohort designs.
- To address biases arising from unequal selection probabilities and participant relatedness.
- To improve the accuracy of genetic association analyses and reduce Type I error inflation.
Main Methods:
- Development of weighted estimators accounting for unequal selection probabilities and differential nonresponse.
- Derivation of variance estimators that incorporate sampling design and potential participant relatedness.
- Analytical and numerical comparisons of proposed weighted estimators against unweighted methods.
- Application to MetaboChip data from the Hispanic Community Health Study/Study of Latinos.
Main Results:
- Weighted estimators provide less biased estimates of genetic association compared to unweighted methods.
- The proposed variance estimators accurately reflect the complex sampling design.
- The methods demonstrate improved Type I error control in genetic association analyses.
- Successful application to real-world genetic data from a large Hispanic/Latino cohort.
Conclusions:
- Weighted estimators are essential for accurate genetic association analysis in complex cohort studies.
- Accounting for sampling design and participant relatedness is crucial for reliable genetic discoveries.
- The developed methods offer a robust approach for analyzing genetic data in population health studies.
- Guidelines and software are provided to facilitate the use of these advanced statistical techniques.