getMME {lme4breeding}

Build the mixed model equations from a lmebreed object

Description

Uses the lmebreed object and builds the coefficient matrix (C) and returns its inverse and the solutions to the mixed model equations.

Usage

getMME(object, vc=NULL, recordsToKeep=NULL)

Arguments

object

an object of class lmebreed.

vc

optional variance components to force in the mixed model equations. This this to be the outpur of the VarCorr function.

recordsToKeep

a numeric vector of indices to say which records should be kept for forming the mixed model equations and get solutions. This is particularly useful when we want to predict new individuals.

Details

Uses the lmebreed object and builds the coefficient matrix (C) and returns its inverse and the solutions to the mixed model equations. It is internally used by the ranef function when the user wants standard errors for the BLUPs.

Value

$Ci

inverse of the coefficient matrix.

$bu

solutions to the mixed model equations

References

Giovanny Covarrubias-Pazaran (2024). lme4breeding: enabling genetic evaluation in the age of genomic data. To be submitted to Bioinformatics.

Douglas Bates, Martin Maechler, Ben Bolker, Steve Walker (2015). Fitting Linear Mixed-Effects Models Using lme4. Journal of Statistical Software, 67(1), 1-48.

See Also

lmebreed– the core function of the package

Examples

data(DT_example)
DT <- DT_example
A <- A_example

ansMain <- lmebreed(Yield ~ Env + (1|Name),
                    relmat = list(Name = A ),
                    data=DT)
mme <- getMME(ansMain)



##########################################
## showing how to predict the individuals
## that didn't have records in the dataset
##########################################
data(DT_cpdata)
DT <- DT_cpdata
GT <- GT_cpdata
MP <- MP_cpdata
#### create the variance-covariance matrix
A <- A.mat(GT) # additive relationship matrix
A <- A + diag(1e-4, ncol(A), ncol(A))
#### look at the data and fit the model
head(DT)
mix1 <- lmebreed(Yield~ (1|id) + (1|Rowf) + (1|Colf),
                 relmat=list(id=A),
                 control = lmerControl(
                   check.nobs.vs.nlev = "ignore",
                   check.nobs.vs.rankZ = "ignore",
                   check.nobs.vs.nRE="ignore"
                 ),
                 data=DT)

vc <- VarCorr(mix1); print(vc,comp=c("Variance"))
# the new dataset includes more individuals
DT2 <- DT
DT2$Yield <- imputev(DT2$Yield)
mix1expanded <- lmebreed(Yield~ (1|id) + (1|Rowf) + (1|Colf),
                 relmat=list(id=A),
                 control = lmerControl(
                   check.nobs.vs.nlev = "ignore",
                   check.nobs.vs.rankZ = "ignore",
                   check.nobs.vs.nRE="ignore",
                   calc.derivs=TRUE,
                   optCtrl=list(maxeval=1)
                 ),
                 data=DT2)

vc <- VarCorr(mix1expanded); print(vc,comp=c("Variance"))
# predict the individuals that didn't have records in the dataset
res <- getMME(object=mix1expanded, vc=VarCorr(mix1), recordsToKeep = which(!is.na(DT$Yield)) )