Submitted to: National Cotton Council Beltwide Cotton Conference
Publication Type: Abstract Only
Publication Acceptance Date: January 6, 2003
Publication Date: June 1, 2003
Citation: WU, J., JENKINS, J.N., MCCARTY JR, J.C., ZHU, J. APPLICATIONS OF MIXED LINEAR MODEL APPROACHES ON COTTON QUANTITATIVE GENETICS. CD-ROM. Proceedings Beltwide Cotton Conference. 2003. p. 825. Technical Abstract: Many practical genetic models have been developed based on the idea of analysis of variance (ANOVA) proposed by Fisher (1925). These genetic models include NC I, NC II designs, and diallele crossed models. The application of the ANOVA approach in quantitative genetic analyses greatly facilitated the development in quantitative genetics. However, the ANOVA approach has some limitations: (1) it can not analyze unbalanced data without bias and (2) it can not analyze some complicated genetic models. The mixed linear model approach developed in the 1970s can overcome the limitations of ANOVA. Cockerham (1980) proposed the generalized genetic model which can be used to extend to other complicated genetic models. Zhu has developed several practical genetic models based on Cockerham's (1980) idea (Zhu 1992, 1993a, b, 1994, 1996; Zhu and Weir 1994a,b). Mixed linear model approaches were used to estimate the genetic variance and covariance components and predict the genetic effects. The applications of mixed linear model approaches in cotton quantitative traits have included the following aspects: (1) diallel model analyses, (2) seed model analyses, (3) prediction of genotype values in advance, (4) developmental trait analyses, (5) quantitative trait loci (QTL) analyses, and (6) complex trait analyses. The corresponding programs are available and downloadable at web site http://msa.ars.usda.gov/ms/msstate/csrl/jenkins,.htm.