A haplotype-based test of association using data from cohort and nested case-control epidemiologic studies

Area of Science:

• Genetics
• Biostatistics
• Epidemiology

Background:

• Haplotype-based risk models are crucial for disease-gene association studies.
• Ambiguity in haplotype status from genotype data poses challenges in population studies.
• Existing statistical tests are limited to retrospective or cross-sectional designs.

Purpose of the Study:

• To develop a novel statistical test for detecting haplotype-based disease associations in prospective cohort and nested case-control studies.
• To extend association testing methods to handle genotype data directly, accommodating potential ambiguities.
• To incorporate time-to-event data and adjust for covariates in genetic association analyses.

Main Methods:

• Developed a score test based on partial likelihood under a proportional hazard model for haplotype effects.
• Derived an induced hazard function from genotype data to analyze prospective study designs.
• The method accounts for differential follow-up, time-dependent covariates, and age-at-onset information.

Main Results:

• The proposed test statistic is valid for small sample sizes.
• The test demonstrates power in detecting associations, even with missing genotype data.
• Simulations using GPX1 and GPX3 genomic regions validated the approach.

Conclusions:

• The developed statistical test provides a robust method for genetic association studies in prospective designs.
• It effectively utilizes genotype data and handles complexities like missingness and time-dependent factors.
• This approach enhances the power and accuracy of identifying disease-related genomic regions.