Genetic association tests in family samples for multi-category phenotypes

Area of Science:

• Human genetics
• Statistical genomics
• Bioinformatics

Background:

• Recent advances in sequencing technology and statistical methods have accelerated discoveries in human genetics.
• Existing statistical methods primarily focus on continuous or binary traits, leaving a gap for analyzing multinomial categorical traits in family samples.
• There is a need for efficient statistical tools to analyze complex, multi-category phenotypes in genetic studies involving related individuals.

Purpose of the Study:

• To propose and evaluate efficient statistical tests for analyzing multinomial and ordinal categorical traits in family-based genetic association studies.
• To address the lack of suitable methods for complex phenotype analysis in related samples.
• To provide a reliable computational tool for large-scale genetic studies.

Main Methods:

• Development of an efficient score test and an alternative Wald statistic for analyzing multinomial traits in family samples.
• Extension of the methodology to accommodate ordinal traits.
• Extensive simulation studies to assess type-I error rates and statistical power compared to existing methods like multinomial logistic regression.

Main Results:

• The proposed score and Wald tests demonstrated well-controlled type-I error rates in simulations.
• Multinomial logistic regression showed an inflated type-I error rate when applied to family samples.
• The score test was applied to the Framingham Heart Study to identify genetic variants associated with diabesity, a multi-category phenotype.

Conclusions:

• The proposed score and Wald tests offer correct type-I error rates and comparable statistical power for analyzing categorical traits in family studies.
• The score test is more computationally efficient than the Wald test for large-scale genetic association studies.
• Computer implementations are available for analyzing both multinomial and ordinal traits, facilitating their application in genetic research.