Skip to content
/ ezCV Public

Cross-validation without doing cross validation

License

Notifications You must be signed in to change notification settings

camult/ezCV

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

13 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

GBLUP_CV: GBLUP and Leave-one-out Cross-validation.

Description

This function solves univariate linear mixed models by likelihood methods. The optimization methods is based on Efficient Mixed Model Association (Kang et al. 2008).

Usage

GBLUP_CV(formula, data, K, CV = TRUE, folds = NULL, weights = NULL)

Arguments

Argument Description
formula an object of class "formula" (or one that can be coerced to that class): a symbolic description of the model to be fitted.
data an optional data frame containing the variables in the model.
K Covariance matrix for random effects.
CV (logical) if TRUE the Leave-one-out Cross-validation is performed, default TRUE.
folds is a vector with validation sets used to perform k-folds cross-validation. Should be NULL or a numeric vector. If it is NULL, LOOCV will be performed.
weights is an optional vector to allow for heterogeneous error variance: $Var[\varepsilon_i] = R_i \sigma^2_e$ . Should be NULL or a numeric vector. The length of weights must be equal to the number of individuals.

Value

A list with genetic and residual variances; a vector with BLUPs for random effects; a vector for BLUEs of fixed effects; the log-likelihood. AIC and BIC statistics. If CV is TRUE, a vector with BLUPs for random effects using LOOCV method.

References

Hyun Min Kang, Noah A. Zaitlen, Claire M. Wade, Andrew Kirby, David Heckerman, Mark J. Daly and Eleazar Eskin, 2008. Efficient control of population structure in model organism association mapping. Genetics 178:1709-1723. doi:10.1534/genetics.107.080101.

Gianola D, Schon C-C, 2016. Cross-Validation Without Doing Cross-Validation in Genome-Enabled Prediction. G3: Genes|Genomes|Genetics. 6(10):3107-3128. doi:10.1534/g3.116.033381.

Examples

 ## Not to run ##
 
 ## GBLUP_CV(Phen ~ Effect, data=Data, K=G)
 
 ## End(Not run)
 

cvBGBLUP: Cross-Validation Without Doing Cross-Validation on Bayesian GBLUP

Description

Performs a Leave-One-Out Cross-Validation Without Doing Cross-Validation using Gibbs Sampling results.

Usage

cvBGBLUP(n, Y, g, K, Vg, Ve)

Arguments

Argument Description
n is the number of observations. If length(n) > 1, the length is taken to be the number required.
Y is a phenotypic matrix with n rows and 1 column.
g is vector with the BLUP solution for the genetic values.
K is a relationship matrix with m rows and m columns.
Vg is the posterior mean of the genetic variance.
Ve is the posterior mean of the residual variance.

Examples

 ## Not to run ##
 
 ## cvBGBLUP(n=nSamp, Y=y, g=gHat, K=G, Vg=gVAR, Ve=eVAR)
 
 ## End(Not run)
 

RRBLUP_CV: Ridge Regression, Leave-one-out and Leave-D-out Cross-validation

Description

This function solves univariate linear mixed models by likelihood methods. The optimization methods is based on Efficient Mixed Model Association (Kang et al. 2008).

Usage

RRBLUP_CV(formula, data, Z, CV = TRUE, folds = NULL, weights = NULL)

Arguments

Argument Description
formula an object of class "formula" (or one that can be coerced to that class): a symbolic description of the model to be fitted.
data an optional data frame containing the variables in the model.
Z SNP markers matrix with animals in rows and markers in columns.
CV (logical) if TRUE the Leave-one-out Cross-validation is performed, default TRUE.
folds is a vector with validation sets used to perform k-folds cross-validation. Should be NULL or a numeric vector. If it is NULL, LOOCV will be performed.
weights is an optional vector to allow for heterogeneous error variance: $Var[\varepsilon_i] = R_i \sigma^2_e$ . Should be NULL or a numeric vector. The length of weights must be equal to the number of individuals.

Value

A list with genetic and residual variances; a vector with BLUPs for random effects; a vector for BLUEs of fixed effects; the log-likelihood. AIC and BIC statistics. If CV is TRUE, a vector with BLUPs for random effects using LOOCV method.

References

Hyun Min Kang, Noah A. Zaitlen, Claire M. Wade, Andrew Kirby, David Heckerman, Mark J. Daly and Eleazar Eskin, 2008. Efficient control of population structure in model organism association mapping. Genetics 178:1709-1723. doi:10.1534/genetics.107.080101.

Gianola D, Schon C-C, 2016. Cross-Validation Without Doing Cross-Validation in Genome-Enabled Prediction. G3: Genes|Genomes|Genetics. 6(10):3107-3128. doi:10.1534/g3.116.033381.

Examples

 ## Not to run ##
 
 ## RRBLUP_CV(Phen ~ Effect, data=Data, Z=M)
 
 ## End(Not run)
 

cvBayes: Cross-Validation Without Doing Cross-Validation on Gibbs Sampling

Description

Performs a Leave-One-Out Cross-Validation Without Doing Cross-Validation using Gibbs Sampling results.

Usage

cvBayes(Y, B, X, varE)

Arguments

Argument Description
Y is a phenotypic matrix with n rows and 1 column.
B is a matrix with the posterior distribution of markers effect.
X is a matrix with markers in columns and animals in rows.
varE is a vector with the posterior distribution of residual variance.

Examples

 ## Not to run ##
 
 ## cvBayes(Y=y, B=Effect, X=X, varE=BA_varE)
 
 ## End(Not run)
 

MME: Mixed Model Equation

Description

Solves a univariate mixed model of form $y=X\beta+Zu+e$ .

Usage

MME(y, X, Z, K)

Arguments

Argument Description
y a matrix with n rows and 1 column.
X a matrix with n rows and x columns.
Z a matrix with n rows and m columns.
K a matrix with m rows and m columns.

Examples

 ## Not to run ##
 
 ## MME(Y, X, Z, K)
 
 ## End(Not run)
 

LOOCV_DG: Cross-Validation Without Doing Cross-Validation

Description

Performs a Leave-One-Out Cross-Validation Without Doing Cross-Validation using a relationship matrix.

Usage

LOOCV_DG(y, K, lambda, g)

Arguments

Argument Description
y is a phenotypic matrix with n rows and 1 column.
K is a relationship matrix with m rows and m columns.
lambda is the ratio bwtween residual and genetic variance.
g is the BLUP solution for the genetic values.

Examples

 ## Not to run ##
 
 ## LOOCV_DG(Y, G, lambda)
 
 ## End(Not run)
 

kCV_DG: K-folds Cross-Validation Without Doing Cross-Validation

Description

Performs a K-folds Cross-Validation Without Doing Cross-Validation using a relationship matrix.

Usage

kCV_DG(Y, K, lambda, folds, g)

Arguments

Argument Description
Y is a phenotypic matrix with n rows and 1 column.
K is a relationship matrix with m rows and m columns.
lambda is the ratio bwtween residual and genetic variance.
folds is a vector with validation sets used to perform k-folds cross-validation.
g is the BLUP solution for the genetic values.

Examples

 ## Not to run ##
 
 ## kCV_DG(Y, K, lambda, folds)
 
 ## End(Not run)
 

RRBLUP: Ridge Regression BLUP

Description

Solves a univariate mixed model of form $y=X\beta+Mu+e$

Usage

RRBLUP(y, X, M)

Arguments

Argument Description
y a matrix with n rows and 1 column
X a matrix with n rows and x columns
M a matrix with n rows and m columns

Examples

 ## Not to run ##
 
 ## RRBLUP(y, X, M)
 
 ## End(Not run)
 

LOOrrDG: rrBLUP Cross-Validation Without Doing Cross-Validation

Description

Performs a Leave-One-Out Cross-Validation Without Doing Cross-Validation using a relationship matrix.

Usage

LOOrrDG(Y, X, B, lambda)

Arguments

Argument Description
Y is a phenotypic matrix with n rows and 1 column.
X is a matrix with markers in columns and animals in rows.
B is matrix with markers effect.
lambda is the ratio bwtween residual and genetic variance.

Examples

 ## Not to run ##
 
 ## LOOrrDG(Y, X, B, lambda)
 
 ## End(Not run)
 

About

Cross-validation without doing cross validation

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published