Bayesian mapping of genomewide interacting quantitative trait loci for ordinal traits

Area of Science:

• Quantitative genetics
• Statistical genomics
• Bioinformatics

Background:

• Mapping interacting quantitative trait loci (QTL) is crucial for understanding complex traits.
• Previous Bayesian frameworks focused on continuous traits, limiting analysis of ordinal phenotypes.
• Ordinal traits, common in biological studies, require specialized statistical approaches for QTL mapping.

Purpose of the Study:

• To extend the composite model space approach for mapping interacting QTL to complex ordinal traits.
• To develop a Bayesian statistical framework for jointly modeling QTL and environmental effects on ordinal phenotypes.
• To provide a flexible and convenient method for genomewide interacting QTL analysis of ordinal traits in experimental crosses.

Main Methods:

• Utilized an ordinal probit model (threshold model) assuming an underlying latent continuous trait.
• Developed a data augmentation approach for jointly generating latent data and thresholds.
• Integrated the ordinal probit model with the composite model space framework for QTL mapping.

Main Results:

• Successfully detected novel QTL and epistatic effects for an ordinal trait (dead fetuses) in a mouse F(2) intercross.
• Demonstrated the method's utility and flexibility using a simulated dataset.
• Implemented the developed methodology in the freely available R/qtlbim package.

Conclusions:

• The proposed Bayesian ordinal probit model combined with the composite model space approach enables genomewide interacting QTL analysis for ordinal traits.
• The R/qtlbim package facilitates the application of this Bayesian methodology for continuous, binary, and ordinal traits.
• This advancement significantly enhances the capability for dissecting genetic architectures of complex ordinal traits in experimental populations.