Post-GWAS: where next? More samples, more SNPs or more biology?

Area of Science:

• Quantitative genetics
• Population genetics
• Genomics

Background:

• Foundational assumptions of genome-wide association studies (GWAS), including additive genetic variation, large effect sizes, and intermediate allele frequencies, are increasingly questioned.
• Nonlinear molecular networks, infinitesimal models of effect sizes, and low-frequency causal polymorphisms challenge traditional GWAS frameworks.

Purpose of the Study:

• To review recent findings that challenge the core assumptions of genome-wide association studies (GWAS).
• To propose a future roadmap for quantitative genetics, emphasizing the integration of population genetic models and molecular biological knowledge.
• To advocate for moving GWAS beyond purely statistical approaches.

Main Methods:

• Review of existing literature and findings questioning GWAS assumptions.
• Proposal of a hybrid approach combining population genetics and molecular biology.
• Application of Bayesian analysis and approximate Bayesian computation for complex model fitting and inference.

Main Results:

• Evidence suggests that genetic variation may not always be additive due to nonlinear molecular networks.
• Relatedness-based analyses support an infinitesimal model for effect sizes, contrasting with the large effects often assumed in GWAS.
• Selection pressures may favor low-frequency, large-effect causal polymorphisms, challenging their detectability in standard GWAS.

Conclusions:

• The optimism surrounding current genome-wide association studies (GWAS) may be overstated due to questioned foundational assumptions.
• Future quantitative genetics research should integrate population genetic models with molecular data for a more comprehensive understanding.
• Bayesian methods, particularly approximate Bayesian computation, offer a robust framework for analyzing complex genetic data and augmenting GWAS in human and agricultural research.