Flexible design for following up positive findings

Area of Science:

• Genetics
• Biostatistics
• Epidemiology

Background:

• Population-based studies increasingly link genetic variants to disease risk.
• Effective follow-up studies are crucial for confirming initial genetic association findings.
• Current methods may not adequately address biases in selecting markers for follow-up.

Purpose of the Study:

• To propose improved sample size calculation methods for follow-up genetic studies.
• To address and correct for regression to the mean (winner's curse) in marker selection.
• To enhance the reliability and power of genetic association studies.

Main Methods:

• Utilized flexible designs from randomized clinical trials for sample size calculations.
• Applied a bootstrap procedure to correct for regression to the mean.
• Developed adaptive sample size adjustments using stage I study results.

Main Results:

• The proposed bootstrap method provides more appropriate sample size calculations than conventional approaches.
• Adaptive use of stage I data maintains nominal global type I error rates.
• Joint analysis of stage I and stage II data is generally more powerful than replication analysis alone.

Conclusions:

• The proposed adaptive sample size calculation method enhances the design of follow-up genetic studies.
• Correcting for the winner's curse is essential for accurate confirmation of genetic associations.
• Flexible designs and joint analysis offer increased power for detecting disease risk variants.