Circumventing multiple testing: a multilocus Monte Carlo approach to testing for association

Area of Science:

• Genetics
• Statistical genetics
• Bioinformatics

Background:

• Dense marker maps are now feasible due to technological advancements.
• Performing individual association tests for each marker leads to multiple testing issues.
• The Bonferroni correction is a common but potentially underpowered solution for multiple testing in dense marker scenarios.

Purpose of the Study:

• To propose a Monte Carlo procedure for global significance assessment in family-based association studies.
• To evaluate the power of the Monte Carlo procedure compared to the Bonferroni correction, especially with linked markers.

Main Methods:

• Computer simulations were used to study the transmission/disequilibrium test (TDT).
• The study examined scenarios with tightly linked markers and varying degrees of association between them.
• The power of the TDT was compared using both Bonferroni and Monte Carlo significance assessment procedures.

Main Results:

• Both Bonferroni and Monte Carlo procedures showed similar power when marker alleles were not associated.
• The Monte Carlo procedure demonstrated significantly higher power than the Bonferroni correction when alleles at linked markers were associated.
• The proposed Monte Carlo method provides a more powerful alternative for genetic association studies with linked markers.

Conclusions:

• The Monte Carlo procedure is recommended for association studies involving potentially associated linked markers.
• This method ensures appropriate significance levels and optimizes statistical power.
• It serves as a robust approach for genetic association analysis in complex marker datasets.