Fast permutation tests and related methods, for association between rare variants and binary outcomes.

Area of Science:

• Genetics
• Statistical genetics
• Bioinformatics

Background:

• Large-scale genetic studies aim to link genetic variants to disease outcomes.
• Rare genetic variants present challenges for standard statistical tests, particularly with unbalanced case-control ratios.
• Existing methods may inadequately control type I error rates for rare variants.

Purpose of the Study:

• To propose novel permutation and approximate unconditional tests for association analysis with rare genetic variants.
• To improve the accuracy and reliability of statistical tests in large genetic studies.
• To address the limitations of standard asymptotic tests in detecting rare variant associations.

Main Methods:

• Development of analytical calculations to approximate the true type I error rate.
• Implementation of permutation tests for rare variant association.
• Application of approximate unconditional tests for genetic association studies.
• Evaluation of methods using numerical simulations and a real case-control dataset.

Main Results:

• The proposed tests significantly improve upon standard testing methods for rare variant association.
• Novel analytical calculations efficiently approximate type I error rates.
• The methods demonstrate enhanced performance, especially in unbalanced study designs.

Conclusions:

• Permutation and approximate unconditional tests provide a robust approach for rare variant association studies.
• These methods offer better type I error control compared to standard techniques.
• The findings are applicable to genetic studies investigating disease causes and drug side effects.