A multipoint method for meta-analysis of genetic association studies

Area of Science:

• Genetics
• Biostatistics
• Population Genetics

Background:

• Current meta-analyses typically examine one genetic polymorphism at a time.
• This approach overlooks information from studies reporting results for multiple, partially overlapping polymorphisms.
• Linkage disequilibrium (LD) between polymorphisms is often not fully utilized in meta-analyses.

Purpose of the Study:

• To present a novel multipoint (multilocus) method for multivariate meta-analysis of population-based case-control genetic association studies.
• To enable multivariate modeling of log(odds ratios) for multiple polymorphisms in linkage disequilibrium (LD).
• To provide a unified framework applicable to both discrete and continuous outcomes.

Main Methods:

• Extension of the general multivariate meta-analysis method.
• Multivariate modeling of log(odds ratios) for polymorphisms in LD.
• Analytical calculation of within-study covariances for linked polymorphisms.
• Genetic model-free approach, adaptable to specific genetic models.
• Unified framework for discrete (odds ratios) and continuous (mean difference) outcomes.

Main Results:

• The method analytically calculates covariances between estimates from linked polymorphisms.
• Requires only an external estimate of pairwise LD (e.g., from HapMap or published studies).
• Demonstrated robustness and utility in two published meta-analyses.
• The method is simple, fast, and extensible to numerous polymorphisms.

Conclusions:

• The developed multipoint multivariate meta-analysis method offers a more comprehensive approach to analyzing genetic association studies.
• It effectively incorporates information from multiple linked polymorphisms, enhancing analytical power.
• The method's simplicity, speed, and broad applicability are expected to lead to its widespread adoption in genetic research.