Impact of Haplotype Frequency Estimates on Likelihood Ratio Statistics
Author Information
Author(s): Curtis David, Xu Ke
Primary Institution: Queen Mary's School of Medicine and Dentistry, London
Hypothesis
Minor differences in haplotype frequency estimates can lead to significant variations in likelihood ratio statistics.
Conclusion
Small differences in haplotype frequency estimates can result in very large differences in likelihood ratio statistics and associated p-values.
Supporting Evidence
- MLOCUS produced a p value of 1.8*10-15.
- GENECOUNTING produced a p value of 5.4*10-4.
- Permutation testing produced a p value of 2.8*10-4.
- The study involved 503 subjects with heroin dependence and 336 controls.
Takeaway
If we guess the frequency of certain gene variations just a little bit wrong, it can make a big difference in the results we get when testing for diseases.
Methodology
Haplotype frequencies were estimated using MLOCUS and GENECOUNTING in subjects with heroin dependence and controls, with permutation testing for significance.
Potential Biases
Potential bias due to reliance on estimated rather than observed haplotype frequencies.
Limitations
The study highlights issues with estimating haplotype frequencies that may not reflect true values.
Participant Demographics
503 subjects with heroin dependence and 336 controls.
Statistical Information
P-Value
1.8*10-15, 5.4*10-4, 2.8*10-4
Statistical Significance
p<0.05
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website