Power of MDR and GMDR for Interaction Detection
Author Information
Author(s): Chen Guo-Bo, Xu Yi, Xu Hai-Ming, Li Ming D., Zhu Jun, Lou Xiang-Yang
Primary Institution: Institute of Bioinformatics, Zhejiang University
Hypothesis
The study investigates the performance of the multifactor dimensionality reduction (MDR) and generalized MDR (GMDR) methods for detecting gene-gene interactions under various experimental scenarios.
Conclusion
The GMDR method with covariate adjustment performs better than the MDR method, especially with a sample size of 1000 to 2000 for detecting gene-gene interactions.
Supporting Evidence
- The GMDR method had a power of over 80% with a sample size of 2000 when the theoretical accuracy was between 0.56 and 0.62.
- When the accuracy was below 0.56, a sample size of at least 4000 was required for sufficient power.
- The study provided empirical guidelines for planning studies to detect gene-gene interactions.
Takeaway
This study helps scientists understand how to better find interactions between genes that affect traits, showing that using the right methods and enough people in a study can make a big difference.
Methodology
The study used simulations to assess the statistical power of GMDR and MDR methods in detecting gene-gene interactions in case-control designs.
Limitations
The study primarily focused on case-control designs and may not be directly applicable to other study designs.
Participant Demographics
The study involved unrelated subjects in a balanced case-control design.
Statistical Information
P-Value
p<0.001
Statistical Significance
p<0.05
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website