SI-SAS is a code written in SAS to produce several selection indexes. To run the code you have to have SAS 9.2 or higher installed in your computer. The first task is to change the folder where the data file to analyze is located and the name of the file. The data file must be in csv format with columns for variables and rows for observations. Two different designs can be analyzed, complete block design and incomplete block design as single or multiple trials. After you made the changes in the code run it and several windows will guide you through the analysis.
Summary results and individual trial results from the International Late Yellow Hybrid - ILYH, (Mid-altitude/Subtropical Three Way Crosses Yellow Hybrids, With High Concentrations of Provitamin A (especially beta-carotene) - CHTPROA) conducted in 2014.
Summary results and individual trial results from the International Late White Hybrid - ILWH, (Tropical Three Way Crosses White Late Normal Hybrid - TTWCWL) conducted in 2014.
One of the most widely used kernel functions in genomic-enabled prediction is the Gaussian kernel. Usually selection of the bandwidth parameter for kernel regression is based on cross-validation. In this study, we propose a Bayesian method for selecting the bandwidth parameter h of a Gaussian kernel as the mode of its posterior distribution. We present a theory for the Bayesian selection of h in a Transformed Gaussian Kernel (TGK) model and its application in two genomic plant breeding data sets (maize and wheat) that were already predicted using the kernel averaging (KA) method within the context of the Reproducing Kernel Hilbert Spaces’ (RKHS KA). We also compared the prediction accuracy of the proposed method (TGK) with a model that uses a Gaussian kernel (GK) and estimates the bandwidth parameter using restricted maximum likelihood method (GK REML). Results for the wheat data set show that the predictive ability of TGK was on average 3% higher than the predictive ability of model RKHS KA, with TGK showing a smaller Predictive Mean Squared Error (PMSE) than the other two approaches. The advantages of the TGK model over GK REML in terms of PMSE were clear for one trait in nine environments. For the maize data set, the TGK model had slightly better prediction accuracy than methods RKHA KA and GK REML.
The marker × environment interaction (M×E) genomic model can be used to generate predictions for untested individuals and identify genomic regions whose effects are stable across environments and others that show environmental specificity. The objectives of this study were: (1) to extend the M×E interaction model using priors that produce shrinkage and variable selection such as Bayesian Ridge Regression (BRR) and BayesB (BB), respectively, and (2) to evaluate the genomic prediction accuracy of M×E interaction, single-environment, and across-environment models using a multi-parental durum wheat population characterized for grain yield, grain volume weight, thousand-kernel weight and heading date in four environments. Breeding value predictions were generated for two prediction problems: cross-validation problem 1 (CV1) and cross-validation problem 2 (CV2). In general, results showed that the M×E interaction model performed better than the single-environment and across-environment models, in terms of minimization of the model residual variance, for both CV1 and CV2. The improved data-fitting gain over the other models was more evident for thousand-kernel weight and heading date (up to two-fold differences) as compared to grain yield and grain volume weight, which showed more complex genetic bases and smaller single-marker effects. Considering the Bayesian models used, BB showed better overall prediction accuracy than BRR. As proof-of-concept for the M×E interaction model, the major controllers of heading date, Ppd and FT on chromosomes 2A, 2B, and 7A, showed stable effects across environments as well as environment-specific effects. For GY, besides the regions on chromosomes 2B and 7A, additional chromosome regions with large marker effects were detected in all chromosome groups
Multienvironment trials (METs) enable the evaluation of the same genotypes under a v ariety of environments and management conditions. We present META (Multi Environment Trial Analysis), a suite of 31 SAS programs that analyze METs with complete or incomplete block designs, with or without adjustment by a covariate. The entire program is run through a graphical user interface. The program can produce boxplots or histograms for all traits, as well as univariate statistics. It also calculates best linear unbiased estimators (BLUEs) and best linear unbiased predictors for the main response variable and BLUEs for all other traits. For all traits, it calculates variance components by restricted maximum likelihood, least significant difference, coefficient of variation, and broad-sense heritability using PROC MIXED. The program can analyze each location separately, combine the analysis by management conditions, or combine all locations. The flexibility and simplicity of use of this program makes it a valuable tool for analyzing METs in breeding and agronomy. The META program can be used by any researcher who knows only a few fundamental principles of SAS.
In agricultural experimentation, a large number of genotypes are normally tested over a wide range of environments. The occurrence of the genotype (G) X environment (E) interaction (GEI) effect further complicates the selection of superior genoty pes for a target population of environments. In the absence of GEI, the superior genotype in one environment may be regarded as the superior genotype in all, whereas the presence of the GEI confirms particular genotypes being superior in particular environments. Several statistical methods were developed for resolved that kind of problems, for example, Additive main effect and multiplicative interaction analysis (AMMI), Site regression (SREG), Partial least square (PLS), Stability analysis and Factorial regression.
A major objective of biometrical genetics is to explore the nature of gene action in determining quantitative traits. This also includes determination of the number of major genetic factors or genes responsible for the traits. Diallel Mating Designs have been designed to deal with the type of genetic experiments that help assess variability in observed quantitative traits arising from genetic factors, environmental factors, and their interactions. Some Diallel Mating Designs are North Carolina Designs, Line by Tester Designs and Diallel designs. AGD-R is a set of R programs that performs statistical analyses to calculate Diallel, Line by Tester, North Carolina. AGD-R contains a graphical JAVA interface that helps the user to easily choose input files, which analysis to implement, and which variables to analyze.
META-R is a set of R programs that performs statistical analyses to calculate BLUEs, BLUPs, genetic correlations among locations and genetic correlations between variables, broad-sense heritability, and other statistics for breeding trials are given too, in order to make boxplots and histograms. Analyses may be performed by location, across management conditions or across all locations. META-R contains a graphical JAVA interface that helps the user to easily choose input files, which analysis to implement, and which variables to analyze.
We apply a Genomic Selection Index (GSI) to simulated and real data sets with four traits and numerically we compared its efficiency with that of the phenotypic selection index (PSI) using the ratio of the GSI response over the PSI response. In addition, we used two additional criteria to compare the GSI vs PSI efficiency: the ratio of the average of the top 10% of the predicted values of the net genetic merit from one to the next selection cycle for PSI and GSI and the Technow inequality. Results from the simulated data were confirmed by the real data, indicating that GSI was more efficient than PSI in terms of time and that the means of the top 10% of the net genetic merit predicted by GSI were higher than that predicted by PSI. Thus, we concluded that the proposed GSI is an efficient choice when the purpose of a breeding program is to select genotypes using genomic selection.
