Development and application of genomic control methods for genome-wide association studies using non-additive models

Area of Science:

• Genetics
• Statistical genetics
• Bioinformatics

Background:

• Genome-wide association studies (GWAS) are crucial for identifying genes linked to complex traits.
• Population substructure and cryptic relatedness can inflate test statistics, leading to false positive associations.
• Existing genomic control (GC) methods are limited for non-additive models, requiring matched allele frequencies.

Purpose of the Study:

• To extend the genomic control (GC) method for non-additive models in GWAS.
• To enable the use of null markers with arbitrary allele frequencies for GC correction.
• To develop and validate new GC methods for robust association testing.

Main Methods:

• Derived analytical expressions for test statistic inflation in recessive, dominant, and over-dominant models.
• Proposed a method to estimate necessary population parameters.
• Developed a GC method using polynomial approximation of the correction coefficient based on allele frequency.
• Extended GC for genotypic tests when the inheritance model is unknown.
• Implemented methods in the R package GenABEL.

Main Results:

• The proposed GC methods effectively control type 1 error rates in the presence of genetic substructure.
• Analytical expressions revealed dependencies of test statistic inflation on allele frequency and population parameters.
• The extended GC methods accommodate non-additive models and arbitrary allele frequencies of null markers.
• Simulations and real data analyses confirmed the effectiveness of the developed methods.

Conclusions:

• The enhanced GC methods provide robust control of spurious associations in GWAS across various inheritance models.
• These methods broaden the applicability of GC, particularly for complex traits with non-additive genetic architectures.
• The developed R package GenABEL facilitates the application of these advanced statistical genetics tools.