Evaluation of strategies for multi-trait association studies in maize architectural traits

Genomic selection and genome-wide association studies (GWAS) are typically performed univariately on a single trait. However, multi-trait approaches that use information from correlated traits measured within specific plant and animal species are an emerging focus of quantitative genetics. Although promising, this area of study requires further evaluation across a wide variety of genetic architectures and species. Therefore, we studied leaf and inflorensce maize architectural traits that have been shown to be associated with putatively pleiotropic genomic loci. Using the eigenvalues from a principal component analysis (PCA) is a computationally efficient and available alternative to true multivariate GWAS model. Preforming GWAS on the PCs of related traits has the potential to uncover genomic markers with pleiotropic effects. To aid in distinguishing between pleiotropic and non-pleiotropic loci, we are currently developing an analytical pipeline that utilizes the simultaneous application of multivariate (PCs approach) and univariate statistical models to associate genomic markers with these traits. Finally post hoc tests using the p values from each association test will be implemented to obtain a single significance value for each genomic marker. In addition to facilitating the identification of genomic regions that potentially harbor pleiotropic loci, we hypothesize that analysis of this data will help calibrate accurate modeling of genetic and non-genetic sources of trait variability.