Critical Values and Variation in Type I Error in COGA Dataset
Author Information
Author(s): George J Papanicolaou, Cristina M Justice, Illija M Kovac, Alexa JM Sorant, Alexander F Wilson
Primary Institution: Genometrics Section, Inherited Disease Research Branch, NHGRI/NIH, Baltimore, MD, USA
Hypothesis
Does the distribution of p-values depend on chromosomal position for different analysis methods?
Conclusion
The study found that the 5 percentile critical values were generally below the expected 0.05, indicating potential liberal use of certain methods.
Supporting Evidence
- The applied pseudo-trait method was used to determine critical values for regression and linkage analyses.
- Results indicated that the distribution of p-values does not depend on chromosomal position for some analysis methods.
- Significant differences in type I error rates were observed for certain pseudo-traits in linkage analysis.
Takeaway
The researchers looked at how often mistakes happen when testing genes, and they found that some methods might make it too easy to say a result is important.
Methodology
The study used the applied pseudo-trait method to analyze SNP data from the COGA dataset, focusing on type I error rates across chromosomes.
Potential Biases
Potential biases may arise from the selection of pseudo-traits and the specific SNP markers used.
Limitations
The study's findings may not be generalizable due to the specific dataset and methods used.
Participant Demographics
Participants were part of the Collaborative Study on the Genetics of Alcoholism.
Statistical Information
P-Value
0.01
Statistical Significance
p<0.05
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website