Understanding the use of Bayes factor for testing candidate genes

Area of Science:

• Genetics
• Bioinformatics
• Animal Breeding

Background:

• Identifying causal mutations for quantitative trait loci (QTL) is crucial for understanding phenotypic variation.
• Candidate gene selection relies on physiological relevance and QTL proximity, but validation is complicated by linkage disequilibrium.
• Existing methods for candidate gene validation may lose statistical power when simultaneously modeling QTL and candidate gene effects.

Purpose of the Study:

• To propose and evaluate the Bayes factor as a statistical tool for validating candidate genes in genetic studies.
• To present a computational procedure for calculating the Bayes factor between candidate gene and QTL models.
• To compare the power of the Bayes factor approach against classical methods for causal mutation identification.

Main Methods:

• Development of a Bayes factor calculation procedure for comparing candidate gene and QTL models.
• Illustration of the procedure using a simulation study.
• Application of the method to experimental data from an intercross between Iberian and Landrace pigs, analyzing single nucleotide polymorphisms (SNPs) in the leptin receptor (LEPR) gene.

Main Results:

• The Bayes factor procedure was successfully applied to both simulated and real genetic data.
• The proposed method demonstrated higher statistical power in identifying causal mutations compared to traditional approaches.
• Analysis of LEPR SNPs in pigs provided a practical example of the Bayes factor's utility.

Conclusions:

• The Bayes factor offers a more powerful statistical framework for validating candidate genes and identifying causal mutations.
• This approach effectively addresses challenges posed by linkage disequilibrium in genetic analyses.
• The developed procedure provides a valuable tool for genetic research, particularly in animal breeding and quantitative genetics.