Skip to main content
ARS Home » Southeast Area » Mississippi State, Mississippi » Crop Science Research Laboratory » Genetics and Sustainable Agriculture Research » Research » Publications at this Location » Publication #150274
Title: A RESURSIVE APPROACH TO DETECT MULTIVARIABLE CONDITIONAL VARIANCE COMPONENTS AND CONDITIONAL RANDOM EFFECTS

Author
item WU, JIXIANG - MISSISSIPPI STATE UNIV
item WU, DONGFENG - MISSISSIPPI STATE UNIV
item Jenkins, Johnie
item McCarty, Jack

Submitted to: Computational Statistics and Data Analysis
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 8/12/2005
Publication Date: 1/1/2006
Citation: Wu, J., Wu, D., Jenkins, J.N., McCarty Jr., J.C. 2006. A recursive approach to detect multivariable conditional variance components and conditional random effects. Computational Statistics and Data Analysis. 50:285-300.

Interpretive Summary: Complex traits in plants and animals are determined by several component traits. Multivariable conditional analysis in a general mixed linear model is helpful in dissecting the gene expression for the complex trait that result from different effects such as environment, genotype, and genotype x environment interaction. We present a recursive approach for constructing a new random variable that can be used to analyze multivariable conditional components and conditional effects. We used end of season plant maps of cotton to illustrate the use of the new analysis. The model determined the contribution of boll size, boll number, and lint percent individually, and in various combinations, to lint yield. The computer program to run the new model can be downloaded at our website http://msa.ars.usda.gov/ms/msstate/jenkins.htm. Results from this model should provide a better understanding of gene expression for complex traits. It can be applied to data from either plant or animal genetics.

Technical Abstract: A complex trait like crop yield is determined by its several component traits. Multivariable conditional analysis in a general mixed linear model is helpful in dissecting the gene expression for the complex trait from different effects, such as environment, genotype, and genotype environment interaction. In this paper, a recursive approach is presented for constructing a new random vector that can be equivalently used to analyze multivariable conditional variance components and conditional effects. End of season plant mapping data including lint yield and three yield components for nine cultivars of upland cotton (Gossypium hirsutum L.) were used to detect the conditional variance components and conditional effects using this new approach. This approach can help identify genotypes to be used in selection studies.