Improving power and robustness for detecting genetic association with extreme-value sampling design

Area of Science:

• Genetics
• Biostatistics
• Statistical genetics

Background:

• Extreme-value sampling is common in genetic association studies.
• Traditional case-control analysis using logistic regression may lose power by ignoring dose-response relationships.
• Linear regression can have biased parameter estimation and inflated Type I errors.

Purpose of the Study:

• To propose a robust and efficient method for analyzing extreme-value sampling data.
• To account for both biased sampling and potential dose-response relationships.
• To improve the power and accuracy of genetic association studies.

Main Methods:

• Developed a novel statistical approach for extreme-value sampling.
• Incorporated modeling of dose-response relationships.
• Adjusted for biased sampling designs in genetic association analyses.

Main Results:

• The proposed method demonstrated higher statistical power than logistic regression.
• The new approach showed greater robustness than standard linear regression.
• Identified stronger associations for several SNPs with high-density lipoprotein cholesterol (HDL-C) levels.

Conclusions:

• Appropriate modeling of quantitative traits is crucial in extreme-value sampling.
• Adjusting for biased sampling is important when dose-response relationships exist.
• The proposed method offers a more powerful and reliable alternative for genetic association studies.