Improving Statistical Power for Genetic Studies with Errors
Author Information
Author(s): Sandra Barral, Chad Haynes, Millicent Stone, Derek Gordon
Primary Institution: Rockefeller University
Hypothesis
Can the LRTae method improve statistical power in genetic association studies when there are phenotype and/or genotype misclassification errors?
Conclusion
The LRTae method is more powerful for detecting associations in case/control studies when misclassification errors are present.
Supporting Evidence
- The LRTae method showed greater power than the LRTstd method across various significance levels.
- Power gains increased as the significance level became more stringent.
- The method provides asymptotically unbiased estimates of case and control genotype frequencies.
Takeaway
This study created a new method to help scientists find genetic links to diseases even when there are mistakes in the data.
Methodology
A double-sampling procedure was developed using a likelihood ratio test framework and applied to simulated case/control data.
Potential Biases
Potential bias due to misclassification errors in case/control studies.
Limitations
The study did not generate genotype misclassification errors due to limitations in determining permutation p-values.
Participant Demographics
Northern European ancestry Caucasian sample population consisting of 169 individuals affected with AS and 202 unaffected individuals matched on age, gender, and ethnicity.
Statistical Information
P-Value
0.0001
Confidence Interval
null
Statistical Significance
p<0.05
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website