Semi-parametric empirical Bayes factor for genome-wide association studies

Area of Science:

• Genetics
• Statistical Genetics
• Bioinformatics

Background:

• Bayes factor analysis is valuable for identifying susceptibility gene variants in genome-wide association studies (GWASs), balancing false negatives and positives.
• The approximate Bayes factor (ABF) relies on a normal prior for effect sizes, but prior misspecification can compromise its accuracy.
• Accurate incorporation of the probability under the alternative hypothesis is crucial for robust association analysis in GWASs.

Purpose of the Study:

• To propose a novel semi-parametric, empirical Bayes factor (SP-EBF) for enhanced accuracy in GWASs.
• To address the limitations of prior misspecification in existing Bayes factor methods.
• To improve the identification of susceptibility gene variants, particularly those with small effect sizes.

Main Methods:

• Developed a semi-parametric, empirical Bayes factor (SP-EBF) utilizing a non-parametric effect-size distribution.
• Estimated the effect-size distribution directly from the genome-wide association study data.
• Applied the SP-EBF to analyze multiple GWAS datasets.

Main Results:

• Analysis revealed a significant number of single nucleotide polymorphisms (SNPs) with small effect sizes across GWAS datasets.
• The SP-EBF assigned substantially greater statistical significance to these small-effect SNPs compared to the ABF.
• The empirical Bayes factor demonstrated improved performance in detecting associations within the analyzed GWAS data.

Conclusions:

• The proposed SP-EBF method effectively incorporates data-driven effect-size distributions, enhancing Bayes factor analysis in GWASs.
• SP-EBF offers improved accuracy over ABF, especially for detecting SNPs with subtle effects.
• This approach holds significant potential for more reliable identification of genetic variants associated with diseases.