A General Framework for the Evaluation of Genetic Association Studies Using Multiple Marginal Models.

Area of Science:

• Genetics
• Biostatistics
• Computational Biology

Background:

• Genetic association studies are crucial for identifying genes linked to diseases.
• Current methods often analyze genetic loci and endpoints separately, limiting comprehensive analysis.
• Understanding inheritance patterns is key to accurate genetic association.

Purpose of the Study:

• To develop a unified analysis framework for simultaneous inference in genetic association studies.
• To provide a flexible method capable of handling various data types and multiple endpoints.
• To enable the detection of pleiotropic effects while considering modes of inheritance.

Main Methods:

• Formulation of multiple marginal regression models using genotype scores as quantitative variables.
• Avoids explicit formulation of correlations between test statistics.
• Accommodates diverse endpoints: binary, count, quantitative, and time-to-event data.

Main Results:

• The approach allows simultaneous assessment of multiple endpoints of different types.
• Enables detection of pleiotropic effects by considering inheritance modes.
• Facilitates simultaneous analysis of multiple genetic loci.
• Demonstrated flexibility through analysis of various data examples.

Conclusions:

• The proposed simultaneous inference procedure offers a unified and flexible framework for genetic association studies.
• This method enhances the ability to detect complex genetic associations, including pleiotropy.
• The approach is applicable to a wide range of genetic and phenotypic data types.