Related Experiment Videos
Method and computer program for controlling the family-wise alpha rate in gene association studies involving multiple
1Obesity Research Center, St. Luke's/Roosevelt Hospital, Columbia University College of Physicians and Surgeons, New York, New York 10025, USA. DBA8@Columbia.edu
Genetic Epidemiology
|April 2, 1998
Summary
This study introduces a simulation-based method to control Type I error rates in multiple significance testing for gene association studies. The new approach accounts for variable correlations, offering increased statistical power compared to existing methods.
Area of Science:
- Genetics
- Statistical genetics
- Bioinformatics
Background:
- Multiple significance testing is common in gene association studies but often lacks appropriate statistical adjustment.
- Existing methods for adjusting alpha rates may be overly stringent by not considering variable correlations or may only provide composite results.
Purpose of the Study:
- To develop and validate a simulation-based method for multiple significance testing in gene association studies.
- To address the limitations of current alpha adjustment methods by incorporating correlation structures and allowing specific variable analysis.
Main Methods:
- A novel simulation-based approach was developed to maintain the nominal alpha rate while accounting for the correlation structure among variables.
- The method's performance was evaluated through simulations, comparing its power against common alternative approaches.
Main Results:
- The proposed method effectively controls the actual alpha rate at the nominal level, considering inter-variable correlations.
- It demonstrates superior statistical power compared to existing methods, especially with increasing numbers of variables and higher intercorrelations.
- The method shows robustness to non-normality and variance heterogeneity, even with unequal group sizes.
Conclusions:
- This simulation-based method provides a more powerful and accurate approach to multiple significance testing in gene association studies.
- It allows for specific variable testing without further alpha adjustment post-detection, leveraging the closure principle in genetic studies.