Testing Hardy-Weinberg proportions in a frequency-matched case-control genetic association study

Area of Science:

• Genetics
• Statistical genetics
• Population genetics

Background:

• Case-control genetic association studies utilize cases (with disease) and controls (without disease) to identify genetic variants.
• Secondary phenotypes are often collected alongside primary disease data, prompting interest in their genetic associations.
• Hardy-Weinberg proportion (HWP) testing is a standard quality control step to identify genotype errors, typically performed on control samples.

Purpose of the Study:

• To evaluate existing Hardy-Weinberg proportion (HWP) testing methods (LRT and mHWP) in case-control studies, particularly when frequency-matched on a secondary phenotype.
• To address the inflated type I error rates observed with current HWP tests in such scenarios.
• To propose and validate novel HWP testing approaches that account for frequency matching to preserve potentially causal genetic variants.

Main Methods:

• Simulations were conducted to assess type I error rates of existing and proposed HWP testing methods.
• The study extended the likelihood ratio test (LRT) and mixture HWP (mHWP) approaches to incorporate frequency-matching information.
• The proposed methods were applied to a genome-wide association study (GWAS) dataset for lung cancer, frequency-matched on smoking status.

Main Results:

• Existing LRT and mHWP approaches demonstrated inflated type I error rates when studies are frequency-matched on a secondary phenotype.
• This inflation can lead to the erroneous exclusion of potentially causal genetic variants from analysis.
• The proposed extended LRT and mHWP approaches effectively controlled type I error probabilities in simulations and real data.
• The novel approaches successfully retained more single-nucleotide polymorphisms (SNPs) for association analysis in the lung cancer GWAS.

Conclusions:

• Standard HWP tests can be unreliable in frequency-matched case-control studies, leading to loss of valuable genetic data.
• The developed extended LRT and mHWP methods provide a more accurate and robust approach to HWP testing in these specific study designs.
• These improved methods enhance the power of genetic association studies by preserving a greater number of potentially relevant genetic markers.