MetNorm (version 0.1)

NormalizeRUVRandClust: RUV-random method for clustering

Description

Given suitable controls and user input, this function may be used to obtain a normalized metabolomics data matrix suitable for clustering

Usage

NormalizeRUVRandClust(RUVRand, maxIter,  nUpdate=maxIter+1,  lambdaUpdate=TRUE,  p=p,...)

Arguments

RUVRand

Output from NormalizeRUVRand

maxIter

Maximum number of iterations

nUpdate

Update the unwanted variation component every nUpdate iterations

lambdaUpdate

A logical indicating whether the regularization parameter needs to be updated

p

The number of clusters to be used in the k-means clustering

...

Other arguments for kmeans

References

De Livera, A. M., Dias, D. A, De Souza, D., Rupasinghe, T., Pyke, J., Tull, D., Roessner, U., McConville, M., and Speed, T. P. (2012). Normalizing and integrating metabolomics data. Analytical chemistry, 84(24), 10768-76.

De Livera, A.M., Aho-Sysi, M., Jacob, L., Gagnon-Bartch, J., Castillo, S., Simpson, J.A., and Speed, T.P. (2014), Statistical methods for handling unwanted variation in metabolomics data

Examples

data(UV)
Y<-data.matrix(UV[,-c(1:3)]) 

##Empirical controls
IS<-Y[,which(colnames(Y)=="IS")] 
r<-numeric(dim(Y)[2])
for(j in 1:length(r)){
  r[j]<-cor(IS,Y[,j])
}
ctl<-logical(length(r))
ctl[which(r>round(quantile(r,0.7),2))]<-TRUE 

## Not run:   
# ruv<-NormalizeRUVRand(Y=Y,ctl=ctl,k=3) 
# ruvclust<-NormalizeRUVRandClust(RUVRand=ruv,
#                    maxIter=200, 
#                    nUpdate=100,
#                    lambdaUpdate=TRUE,
#                    p=2)
# ruvclustY<-ruvclust$newY
# pairs(princomp(ruvclustY,cor=TRUE)$scores[,c(1:3)],
#       pch=as.numeric(UV[,2]), col=UV[,3],
#       main="RUV random for clustering")
# ## End(Not run)