Optimizing strain selection for association studies under hard cost constraints

Area of Science:

• Quantitative genetics
• Population genetics
• Genomics

Background:

• Quantitative genetics methods are powerful in model organisms and diverse natural populations.
• Phenotyping large strain collections is valuable but can be cost-prohibitive.
• Efficient strategies are needed to optimize experimental power within budget constraints.

Purpose of the Study:

• To evaluate optimal subset selection strategies for genome-wide association studies (GWAS) under budget limitations.
• To compare approaches focusing on cost, genetic diversity, or both simultaneously.
• To introduce and validate the ThriftyMD algorithm for resource-limited GWAS cohort design.

Main Methods:

• Simulation studies across various minor allele frequencies (MAFs) and SNP effect sizes.
• Evaluation of cost-focused, diversity-focused, and combined selection strategies.
• Application of methods to the Hybrid Mouse Diversity Panel (HMDP) data.

Main Results:

• Cost-based selection is most effective at low-to-moderate budgets.
• Diversity-based selection is optimal for rare variants (5-10% MAF) or higher costs.
• The ThriftyMD approach, balancing cost and diversity, outperformed other methods in recovering significant loci and maintaining power on the HMDP.
• ThriftyMD selects strains minimizing genetic distance to unselected strains within a budget.

Conclusions:

• There are inherent trade-offs between cost, diversity, and statistical power in GWAS cohort design.
• The ThriftyMD algorithm offers a robust and versatile solution for optimizing GWAS in resource-limited settings.
• This approach enhances experimental power by strategically selecting representative and cost-effective samples.