asreml.batch: ASReml-R batch analysis

In yzhlinscau/AAfun: ASReml-R Added functions

Description

This function carries out batch analysis for mult-trait with same model and also output heritability etc. in ASReml-R package.

Usage

asreml.batch(data,factorN,traitN,FMod=NULL,RMod=NULL, EMod=NULL,
                           mulT=NULL, mulN=NULL,mulR=NULL,corM=NULL,corMout=FALSE,
                           pformula=NULL,pformula1=NULL,pformula2=NULL,
                           pformula3=NULL,pformula4=NULL,maxit=NULL,
                           ped=NULL,pedinv=NULL,ginverse=NULL)

Arguments

data	aim dataset
factorN	A vector with sites of all factors.
traitN	A vector with sites of all traits.
FMod	Fixed mode.
RMod	Randomed variance structure.
EMod	Error variance structure for multi-trait model.
mulT	Value "T" or "TRUE" for multi-trait model, "F"(default).
mulN	Number of trait for one model in multi-trait analysis,2(default).
mulR	Value "T" or "TRUE" to count corr/error matrix, only works for bitrait, "F"(default).
corM	Value "T" or "TRUE" for corr model, "F"(default).
corMout	Value "T" or "TRUE" to output corr matrix, "F"(default).
pformula	Formula for h2 or corr.
pformula1	Formula for h2 or corr.
pformula2	Formula for h2 or corr.
pformula3	Formula for h2 or corr.
pformula4	Formula for h2 or corr.
maxit	Maximum number of iterations.
ped	Value "T" or "TRUE" for animal model with pedigree, "F"(default)..
pedinv	A G-inverse matrix for pedigree from ASReml.Ainverse().
ginverse	a named list with each component identifying a G-inverse matrix.

Value

factorN	A vector contains column site of factors in dataset, such as c(1:3,5)
traitN	A vector contains column site of aim trait in dataset, such as c(10:15)
FMod	Fixed mode,must start in "y~1+fixed.factor" (single trait) or "cbind(y1,y2)~trait+trait:fixed.factor" (multi-trait, cbind(y1,y2,y3,...))
RMod	Random variance structure, such as "~Fam" or "~us(trait):Fam".
EMod	Error variance structure, such as "~units:us(trait)".
pformula	formula for h2 or corr, such as "h2 ~ 4 * V1/(V1+V2)".
maxit	maximum number of iterations, default is 20.

Author(s)

Yuanzhen Lin <[e-mail]>

References

Yuanzhen Lin. R & ASReml-R Statistics. China Forestry Publishing House. 2016

See Also

Website for instant update: yzhlin-asreml.ys168.com

Examples

library(asreml)
library(AAfun)

##### example I for dataset without pedigree
data(PrSpa)
df<-PrSpa

# exmaple 1.1 for sigle trait model
df1=subset(df,Spacing==3)
asreml.batch(data=df1,factorN=1:5,traitN=c(9:13),
             FMod=y~1+Rep+Plot,RMod=~Fam,
             pformula=h2 ~ 4 * V1/(V1+V2))

# exmaple 1.2 for us model 
asreml.batch(data=df1,factorN=1:5,traitN=c(10:13),
             FMod=cbind(y1,y2)~trait+trait:Rep,
             RMod=~us(trait):Fam,
             EMod=~units:us(trait),
             mulT=TRUE,mulN=2,mulR=TRUE,corMout=T,
             pformula=r.g ~ V2/sqrt(V1*V3),
             pformula1=h2.A ~ 4*V1/(V1+V5),
             pformula2=h2.B ~ 4*V3/sqrt(V3+V7))


# exmaple 1.3 for corr model
asreml.batch(data=df1,factorN=1:5,traitN=c(10:13),
             FMod=cbind(y1,y2,y3)~trait+trait:Rep,
             RMod=~corgh(trait):Fam,
             EMod=~units:us(trait), maxit=30,
             mulT=TRUE,mulN=3,mulR=TRUE,corM=TRUE)

##### example II for dataset with pedigree
data(dfm2);df=dfm2

ped<-df[,1:3]  
pedinv<-asreml.Ainverse(ped)$ginv
df1=subset(df,Spacing==3)

# example 2.1 sigle trait model
asreml.batch(data=df1,factorN=1:6,traitN=c(7:14),
             FMod=y~1+Rep,RMod=~ped(TreeID),
             ped=T,pedinv=pedinv,                                   
             ginverse=list(TreeID=pedinv),
             pformula=h2 ~ V1/(V1+V2))


# exmaple 2.2 us model 
asreml.batch(data=df1,factorN=1:6,traitN=c(10:14),
                    FMod=cbind(y1,y2)~trait+trait:Rep,
                    RMod=~us(trait):ped(TreeID),
                    EMod=~units:us(trait),maxit=40,
                    mulT=TRUE,mulN=2,mulR=TRUE,corMout=F,
                    ped=T,pedinv=pedinv,
                    ginverse=list(TreeID=pedinv),
                    pformula=r.g ~ V2/sqrt(V1*V3),
                    pformula1=h2.A ~ V1/(V1+V5),
                    pformula2=h2.B ~ V3/(V3+V7))


# exmaple 2.3  corr model
asreml.batch(data=df1,factorN=1:6,traitN=c(10:14),
             FMod=cbind(y1,y2)~trait+trait:Rep,
             RMod=~corgh(trait):ped(TreeID),
             EMod=~units:us(trait),maxit=40,
             mulT=TRUE,mulN=2,corM=TRUE,
             ped=T,pedinv=pedinv,
             ginverse=list(TreeID=pedinv)) 
              

yzhlinscau/AAfun documentation built on May 21, 2020, 2:19 p.m.