A novel genome-wide association study method for detecting quantitative trait loci interacting with complex...

Area of Science:

• Plant genetics
• Statistical genomics
• Quantitative trait loci (QTL) analysis

Background:

• Modern association analyses utilize diverse panels, including admixed individuals, presenting statistical challenges.
• Detecting quantitative trait loci (QTLs) requires models that account for population structure and interactions.
• Existing models for population-specific QTL detection often require prior knowledge of population structure, which is frequently unavailable.

Purpose of the Study:

• Introduce two novel statistical models for detecting QTLs interacting with complex population structures.
• Address the challenge of analyzing diverse GWAS panels with intricate population structures.
• Improve the detection accuracy of QTLs with population-specific effects and epistatic interactions.

Main Methods:

• Developed two new models incorporating an interaction term between SNP/haplotype block and genetic background into GWAS models.
• Conducted simulation studies comparing proposed models with state-of-the-art methods under various QTL-background interaction levels.
• Applied the models to a real-world soybean dataset to identify associated QTLs.

Main Results:

• The proposed models outperformed classical models in detecting QTLs interacting with polygenes.
• Models that matched simulation settings were most effective for detecting corresponding QTLs.
• One of the new models identified putative associated QTLs in soybean that conventional models missed.

Conclusions:

• The novel models effectively detect QTLs interacting with complex population structures.
• These models enhance the ability to uncover the genetic architecture of complex traits in diverse populations.
• The RAINBOWR package provides implementation of these advanced statistical models for broader use.