Analyses a line-by-tester trial by REML (R.W. Payne).
Options
PRINT = string tokens |
Specifies the output to be produced (model, components, effects, means, monitoring, vcovariance, deviance, Waldtests, missingvalues, covariancemodels, aic, sic, bic, combinability, tests); default mode, comp, wald, comb, test |
|---|---|
PRECOVERY = string tokens |
Controls what summary output is produced about the models that are tried during recovery (deviance, aic, bic, sic, dffixed, dfrandom, change, exit, best); default devi, aic, sic, dfra, best |
LINES = factor |
Specifies the line (usually female parent); no default (must be specified) |
TESTERS = factor |
Specifies the tester (usually male parent); no default (must be specified) |
CONTROLS = factor |
Distinguishes between control and test (line × tester) genotypes; default is that there are no controls |
FIXED = formula |
Fixed model terms, in addition to the TESTERS main effect and any control comparisons; default * i.e. none |
RANDOM = formula |
Random model terms, in addition to the terms involving LINES, TESTERS and EXPERIMENTS that are included automatically; default * i.e. none |
CONSTANT = string token |
How to treat the constant term (estimate, omit); default esti |
FACTORIAL = scalar |
Limit on the number of factors or covariates in each fixed term; default 3 |
EXPERIMENTS = factor |
Specifies the different experiments for a REML meta analysis; default is that the data are all from a single experiment |
PCOMBINABILITYTERMS = formula |
Terms whose combinability effects are to be printed, selected from LINES, LINES.TESTERS and their interactions with EXPERIMENTS; default is to print all of them |
PTERMS = formula |
Terms (fixed or random) for which effects or means are to be printed; default * implies all the fixed terms |
PSE = string token |
Standard errors to be printed with tables of effects and means (differences, estimates, alldifferences, allestimates, none); default diff |
MVINCLUDE = string tokens |
Whether to include units with missing values in the explanatory factors and variates and/or the y-variates (explanatory, yvariate); default * i.e. omit units with missing values in either explanatory factors or variates or y-variates |
RECOVER = string token |
Whether to try to recover with a simpler random model if REML cannot fit the model (yes, no); default no |
METHOD = string token |
How to choose the best model during recovery (aic, sic, bic); default sic |
Parameters
Y = variates |
Response variates |
|---|---|
COMBINABILITY = pointers |
Pointer to tables of combinability effects for each y-variate |
SECOMBINABILITY = pointers |
Pointer to tables of standard errors of combinability effects for each y-variate |
DEVIANCES = variates |
Saves deviances for LINES, LINES.TESTERS and their interactions with EXPERIMENTS |
EXIT = scalars |
Exit status for each y-variate (zero to indicate that the analysis was successful) |
SAVE = REML save structures |
Save structure from the analysis of each y-variate |
Description
VLINEBYTESTER does a mixed-model analysis of data from a line-by-tester trial, using REML. These are trials in which several “lines” (usually female parents) are all mated with a smaller number of testers (usually male parents). Generally, all combinations of parent will be present in the trial, but incomplete arrangements, such as diallels, are also possible. Control, or check, genotypes may also be present.
The lines must be specified (in a factor) by the LINES option. Similarly, the testers must be specified (again in a factor) by the TESTERS option. If there are any control genotypes, these are specified by using the CONTROLS option to supply a factor with a different level for each control genotype, and a single level for all the line-by-tester genotypes.
You can use the FIXED and RANDOM options to specify fixed and random terms to be fitted in the analysis. The FACTORIAL option sets a limit on the number of factors and variates allowed in each fixed term (default), and the CONSTANT option can be set to omit the constant.
By default, the testers main effect is treated as fixed, so it will be added to any fixed model that you specify. However, you can put TESTERS into the random model (specified by RANDOM), if you would prefer it to be treated as random. Likewise, the main effect of the CONTROLS factor (comparisons amongst control genotypes, and between controls and the mean of the tests) is treated as fixed, but can be put into the random model if you prefer. The model terms LINES and LINES.TESTERS are treated as random, and so will be added automatically to any random model that is specified by RANDOM. If an EXPERIMENTS factor is specified, its main effect and its interactions with LINES and TESTERS are also added to the random model.
The PRINT option specifies the output to be produced. Its settings are mainly the same as those of the PRINT option of the REML directive. There are extra settings aic and sic (with a synonym bic) to print the Akaike and Schwarz (Bayesian) information coefficients, respectively. There is a setting combinability to print the BLUPs for LINES, LINES.TESTERS and any interations with EXPERIMENTS (within CONTROLS, if specified). By default, with this setting, BLUPs are printed for all these terms. However, you can set the PCOMBINABILITYTERMS option to a model formula specifying exactly which ones you want. Finally, there is an extra setting tests to print deviances for LINES, LINES.TESTERS and their interations with EXPERIMENTS, so that you can assess whether their effects are genuinely present.
The PTERMS option operates as in REML, to specify the terms whose means and effects are printed; the default is all the fixed terms. Likewise, the PSE option controls the type of standard error that is displayed with the means and effects; the default is to give a summary of the standard errors of differences.
The MVINCLUDE option controls whether units with missing values in the explanatory factors and variates and/or the y-variate are included in the analysis, as in the REML directive.
REML may be unable to achieve a successful fit if the model contains more random terms than are actually needed to explain the random variation.(The REML likelihood may be too flat for any clear optimum to be found.) You can guard against having specified an over-complicated random model using the RANDOM option, by setting option RECOVER=yes. VLINEBYTESTER then tries models removing first one term from RANDOM, then two and so on, until successful. Note: it regards a model as successful, if the REML directive returns an exit status of zero (i.e. successful fitting) and there are no bound or aliased variance parameters.
The METHOD option specifies how to choose the random (and spatial) model if there is more than one possible model with the same number of random terms removed:
aic |
uses their Akaike information coefficients, |
|---|---|
sic or bic |
uses their Schwarz (Bayesian) information coefficients (default). |
The PRECOVERY option specifies the summary output to be produced about the models that are fitted during recovery. The settings are mainly the same as those of the VRACCUMULATE procedure (which is used to store and then print details of the analyses). There is an extra setting, best, to print the description of the best model. The default is to print the best description, together with the deviance, the Akaike and Schwarz (Bayesian) information coefficients and the number of degrees, for all the models.
The Y parameter specifies the response variate. The COMBINABILITY parameter can save a pointer to tables containing the combinability BLUPs, requested by the PCOMBINABILITYTERMS option. Similarly, the SECOMBINABILITY parameter can save a pointer to tables containing the standard errors of the BLUPs. The DEVIANCES parameter can save the deviances printed by the test setting of PRINT, in a variate. The EXIT parameter allows you to save a code from REML, giving the “exit status” of the fit (zero if successful). Finally, you can save the REML save structure from the analysis with the best model, using the SAVE parameter.
Options: PRINT, PRECOVERY, LINES, TESTERS, CONTROLS, FIXED, RANDOM, CONSTANT, FACTORIAL, EXPERIMENTS, PCOMBINABILITYTERMS, PTERMS, PSE, MVINCLUDE, RECOVER, METHOD.
Parameters: Y, COMBINABILITY, SECOMBINABILITY, DEVIANCES, EXIT, SAVE.
See also
Directives: REML, VCOMPONENTS, VSTRUCTURE.
Procedure: VALINEBYTESTER, VARECOVER.
Commands for: REML analysis of linear mixed models.
Example
CAPTION 'VLINEBYTESTER example',!t(\
'33 lines x 2 testers and a check genotype,',\
'in a resolvable incomplete block design'); STYLE=meta,plain
SPLOAD [PRINT=*] '%gendir%/examples/VLIN-1.gsh'
VLINEBYTESTER [LINES=Lines; TESTERS=Testers; CONTROLS=Controls;\
RANDOM=Replicates/Blocks] Yield