Estimation of genetic correlations on sweet corn inbred lines using SAS mixed model

• Main Authors: Kashiani, Pedram, Saleh, Ghizan
• Format: Article
• Language: English
• Published: Science Publications 2010
• Online Access: http://psasir.upm.edu.my/id/eprint/14569/1/ajabssp.2010.309.314.pdf; http://psasir.upm.edu.my/id/eprint/14569/; http://www.thescipub.com/abstract/10.3844/ajabssp.2010.309.314

Problem statement:Genetic correlations among traits refer to the extent of relatedness among them due to genetic causes. Estimating genetic correlations for quantitative traits is tedious if done manually. Approach:However, the use of the computer software SAS, applying mixed-model analysis of variance has facilitated many recent studies in evolutionary quantitative genetics. Results:In this two-way statistical model, the variance component corresponding to the random statement is the covariance associated with a level of the random factor across levels of the fix factor. Therefore, the SAS model has a natural application for estimating genetic correlations among traits measured. Correlation studies were undertaken for 10 yield-related traits on a series of near-homozygous sweet corn inbred lines obtained from various tropical source populations. The SAS program was used to estimate genetic correlation coefficients among traits observed, where effects of blocks were considered fixed while effects of inbred lines as random. The "ASYCOV" was added to the "PROC MIXED" statement in order to produce the variance-covariance matrix of variance components. The "TYPE = UN" option requested in "RANDOM" statement resulted in an unstructured covariance matrix for each inbred line being estimated, while the "G" and "GCORR" options produced genetic variance-covariance matrix and genetic correlation matrix between traits, respectively. Results showed that there was no significant difference between genetic correlations estimated by SAS MIXED model and those estimated by manual calculation. Conclusion/Recommendations: This indicated that SAS has the natural capability to estimate genetic correlations among traits measured, as opposed to manual methods employing quantitative genetics equations.