CNVs with highes correlations

Description

Obtains the CNV blocks that have the highes correlation according to rankVariables.

Usage

1
getNamesProbes(x, min.correlation = 0.3)

Arguments

x

data.frame with the ranking of the CNV blocks according to their correlation to the population centroids.

min.correlation

value of the minimum correlation to select the variables from the ranking.

Details

The output of getNamesProbes can be used to re-train a discriminant analysis discriminCNV.

Value

produces a vector of characters with the CNV blocks that are most relevant in the discrimination summarized on x.

Examples

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
## Not run: 

# 'HapMap270reducedData.RData' file can be downloaded from ...

load("HapMap270reducedData.RData")

cp<-discrimin.cnv(mat.f,pop.cla)

var.proj<-rank.variables(cp,cnv.blocks=cnv.blocks)
select<-getNamesProbes(var.proj, min.correlation=0.5)
#re-train discriminant
cp.2<-discrimin.cnv(mat.f, pop.cla, select=select)
plot(cp.2)

#re-compute correspondence analysis with selected variables.
cm<-dudi.acm.fortran(mat.f[select], scan=FALSE, nf=3)
plot(cm)

## End(Not run)