Semiparametric Allelic Tests for Mapping Multiple Phenotypes: Binomial Regression and Mahalanobis Distance

Area of Science:

• Genetics
• Statistical genetics
• Bioinformatics

Background:

• Multiple underlying quantitative precursors often lead to binary phenotypes.
• Genetic variants can exhibit pleiotropy, affecting multiple traits simultaneously.
• Simultaneous analysis of correlated traits can enhance statistical power for genetic association studies.

Purpose of the Study:

• To explore and evaluate novel allelic tests for multivariate association mapping.
• To compare the power of these new allelic methods against existing genotype-level methods.
• To propose a hybrid approach for robust multivariate association testing.

Main Methods:

• Developed and investigated two allelic tests: Binomial regression-based Association of Multivariate Phenotypes (BAMP) and Distance-based Association of Multivariate Phenotypes (DAMP).
• Incorporated both discrete and continuous phenotypes into the allelic models.
• Compared the power of BAMP and DAMP with the genotype-level test MultiPhen using simulations and real data.

Main Results:

• Allelic tests (BAMP and DAMP) demonstrated marginally higher power than MultiPhen for multivariate phenotypes.
• Allelic tests showed substantially greater power for one/two binary traits under a recessive mode of inheritance.
• Real data analysis supported the simulation findings, validating the utility of the proposed methods.

Conclusions:

• Allelic association tests provide a powerful alternative for multivariate phenotype analysis.
• The proposed methods, BAMP and DAMP, enhance the ability to identify genetic factors influencing multiple traits.
• A hybrid approach combining MultiPhen and BAMP based on Hardy-Weinberg Equilibrium (HWE) status is recommended for optimal performance.