Recessive/dominant model: Alternative choice in case-control-based genome-wide association studies

Area of Science:

• Genetics
• Bioinformatics
• Statistical Genetics

Background:

• Case-control genome-wide association studies (GWAS) commonly use additive genetic models.
• Additive models encode genotypes (e.g., AA, Aa, aa) numerically, assuming distinct contributions to phenotype.
• The binary nature of case-control phenotypes suggests alternative models like recessive or dominant may be more suitable.

Purpose of the Study:

• To investigate the feasibility of using recessive/dominant models as alternatives to additive models in case-control GWAS.
• To compare the performance of different genetic models through simulation and real data experiments.

Main Methods:

• Comparative experiments using chi-square tests and logistic regression.
• Simulation experiments to evaluate model performance.
• Real data experiments to assess precision and area under the curve (AUC).

Main Results:

• Simulation experiments showed recessive models outperformed additive models, with a 5% AUC increase and 61% improvement in risk SNP identification.
• Real data experiments indicated recessive/dominant models achieved higher precision and AUC compared to additive models.
• Specifically, AUC for recessive models was 20% higher, and for dominant models, it was 13% higher than additive models.

Conclusions:

• Recessive/dominant genetic models offer a viable alternative to additive models in case-control GWAS.
• These models enhance the accuracy and precision of identifying genetic risk factors.
• The findings suggest a new methodological approach for case-control genetic studies.