Comparing Methods for Identifying Genetic Markers in Disease
Author Information
Author(s): Gu Fangyi, Monsees Genevieve, Kraft Peter
Primary Institution: Harvard School of Public Health
Hypothesis
Can two-locus tests of association that incorporate gene × gene interaction be more powerful than single-locus tests for identifying disease susceptibility loci?
Conclusion
The study found that while two-locus tests were powerful in detecting associations, they lacked specificity in identifying the most promising genetic markers.
Supporting Evidence
- Two-locus tests showed perfect power to detect associations with RA.
- Models containing denseSNP6_3426 and denseSNP6_3436 provided the strongest evidence for association.
- Every candidate SNP was significantly associated with RA at the genome-wide level when tested with certain markers.
Takeaway
The researchers looked at how different methods can help find genes that might cause diseases, and they found that some methods are good at finding genes but not so good at picking the best ones to study further.
Methodology
The study used single- and two-locus tests on 29 SNPs from simulated case-control data, applying likelihood-ratio tests, AIC, and BIC for model selection.
Limitations
The two-locus tests were not very specific, leading to potential misinterpretation of results.
Participant Demographics
1500 cases and 2000 controls from affected sib pairs.
Statistical Information
P-Value
p<1.0 × 10-14
Statistical Significance
p<1.0 × 10-14
Want to read the original?
Access the complete publication on the publisher's website