Meta-analysis of Haplotype-Association Studies
Author Information
Author(s): Bagos Pantelis G
Primary Institution: Department of Computer Science and Biomedical Informatics, University of Central Greece
Hypothesis
What are the optimal methods for performing meta-analysis of haplotype association studies?
Conclusion
The study found that traditional methods for analyzing haplotype associations often yield inflated type I error rates, and many significant results may not hold up under more rigorous analysis.
Supporting Evidence
- The study found that haplotype association methods are more powerful than single marker analyses.
- Approximately half of the significant results in previous studies were refuted upon re-analysis.
- Meta-analyses of haplotypes produce significant results twice as often as those using single nucleotide polymorphisms.
Takeaway
This study looks at how to combine results from different studies about gene variations and their links to diseases, showing that some methods might make it look like there are more connections than there really are.
Methodology
The study used a meta-analysis framework to compare different statistical methods for analyzing haplotype associations, including logistic regression and Poisson regression.
Potential Biases
The methods used in many studies may lead to inflated type I error rates due to multiple comparisons.
Limitations
The study assumes that haplotypes are directly observed, which may not always be the case in real-world applications.
Participant Demographics
The analysis included data from various studies, with a total sample size averaging around 5,018 participants.
Statistical Information
P-Value
p<0.05
Statistical Significance
p<0.05
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website