Two-stage designs for gene-disease association studies

Area of Science:

• Genetics
• Biostatistics
• Bioinformatics

Background:

• Gene-disease association studies are crucial for understanding complex traits.
• Traditional study designs can be resource-intensive, limiting the number of genes evaluated.
• Optimizing cost-efficiency is essential for maximizing the power of genetic association studies.

Purpose of the Study:

• To describe a novel two-stage design for gene-disease association studies.
• To maximize statistical power under a total cost constraint, defined by gene evaluations.
• To provide a cost-effective strategy for genetic marker screening.

Main Methods:

• A two-stage design involving initial screening of all genes on a subset of individuals.
• Subsequent evaluation of promising genes on additional subjects in a second stage.
• Consideration of both independent and correlated gene scenarios.
• Simulation studies to evaluate design performance.

Main Results:

• A two-stage approach effectively eliminates wasted resources on genes unlikely to be associated with disease.
• Utilizing 75% of resources in stage 1 to screen all markers and the remaining 25% to evaluate the top 10% of markers offers near-optimal power.
• This strategy is effective for both independent and small-correlation gene scenarios.
• Stage 1 involves screening all markers on approximately one quarter of the total sample size.

Conclusions:

• The proposed two-stage design offers a powerful and cost-efficient method for gene-disease association studies.
• The general guideline of screening all markers on a subset (approx. 25% of sample size) and re-evaluating top markers (approx. 10%) is robust across various configurations.
• This approach significantly enhances the efficiency of genetic association studies by optimizing resource allocation.