Computation of target G ('knowledge-based constant correlation model').

The $p \times p$ target G is computed from the $n \times p$ data matrix. It is defined as follows ($i,j = 1,...,p$): $$t_{ij}=\left\{ $$$$\begin {array} {ll} $$$$s_{ii}\;&\mbox{if}\;i=j\\ $$$$\bar{r}\sqrt{s_{ii}s_{jj}}\;&\mbox{if}\;i\neq j, i\sim j\\ $$$$0\;&\mbox{otherwise} $$$$\end{array} $$$$\right.$$ where $\bar{r}$ is the average of sample correlations and $s_{ij}$ denotes the entry of the unbiased covariance matrix in row $i$, column $j$. The notation $i\sim j$ means that genes $i$ and $j$ are connected, i.e. genes $i$ and $j$ are in the same gene functional group.

Usage

targetG(x, genegroups)

Arguments

x: A $n \times p$ data matrix.
genegroups: A list of genes obtained using the database KEGG, where each entry itself is a list of pathway names this genes belongs to. If a gene does not belong to any gene functional group, the entry is NA.

Value

A $p \times p$ matrix.

References

J. Schaefer and K. Strimmer, 2005. A shrinkage approach to large-scale covariance matrix estimation and implications for functional genomics. Statist. Appl. Genet. Mol. Biol. 4:32.
M. Jelizarow, V. Guillemot, A. Tenenhaus, K. Strimmer, A.-L. Boulesteix, 2010. Over-optimism in bioinformatics: an illustration. Bioinformatics. Accepted.

Author

Monika Jelizarow and Vincent Guillemot

Examples


# A short example on a toy dataset
# require(SHIP)
data(expl)
attach(expl)
#> The following objects are masked from expl (pos = 3):
#> 
#>     genegroups, x
#> The following objects are masked from expl (pos = 4):
#> 
#>     genegroups, x
#> The following objects are masked from expl (pos = 5):
#> 
#>     genegroups, x
tar <- targetG(x,genegroups)
#> Error in targetG(x, genegroups): could not find function "targetG"
which(tar[upper.tri(tar)]!=0) # not many non zero coefficients !
#> Error in as.vector(x, mode): cannot coerce type 'closure' to vector of type 'any'