% -*- TeX:Rnw:UTF-8 -*-
% ----------------------------------------------------------------
% .R knitr file  ************************************************
% ----------------------------------------------------------------
%%
<<echo=FALSE>>=
###############################################
##                                           ##
## (c) Andrej Blejec (andrej.blejec@nib.si)  ##
##                                           ##
###############################################
@
<<echo=FALSE,results='hide'>>=
options(width=70)
@

\section{Canonical correlation - rCCA}

Dimensionality of the problem requires regularized Canonical correlation analysis (rCCA). We will follow the methods implemented in the package \pkg{mixOmics}. For explanation see \url{http://mixomics.org/methods/rcca/}.

<<>>=
library(mixOmics)
@

Check colinearity of covariance matrices

Metabolites
<<>>=
covM <- cov(M)
det(covM)
heatmap(covM)
@

Transcripts, small submatrix, determinant almost zero
<<>>=
covT <- cov(T[,sample(1:ncol(T), min(50, ncol(T)))])
det(covT)
heatmap(covT)
@