intCor {MergeMaid} | R Documentation |
Given a mergeExpressionSet, this function calculates the study specific correlation matrices, and, for each gene, the correlation of correlations.
intCor(x,method= c("pearson", "spearman"),exact,...)
x |
Object of class mergeExpressionSet. |
method |
Method used to calculate correlation coefficient. If exact is TRUE, the available methods to use is "spearman" and "pearson"; If exact is FALSE, the available methods to use is "pearson". |
exact |
If exact is TRUE, we use the standard method the calculate the integrative correlation; If exact is FALSE, we use the approximate method the calculate. |
... |
Not implemented at this time |
Integrative correlation coefficients are calcualted as follows. The first step is to identify the n genes common to all studies. Within each study, we calculate the correlation coefficient between gene g, and every other common gene. This gives a vector of length n-1. For a pair of studies, S1 and S2, we calculate the correlation of correlations for gene g. When there are more than 2 studies under consideration, all pairwise correlation of correlations are calculated and averaged.
The output is an object of class mergeCor.
if(require(Biobase) & require(MASS)){ data(mergeData) merged <-mergeExprs(sample1,sample2,sample3) corcor <-intCor(merged,method="spearman") plot(merged) hist(corcor) corcor <-intCor(merged,method="pearson",exact=FALSE) corcor <-intCor(merged[1:2]) corcor <-intCor(merged,exact=TRUE) vv<-c(1,3) corcor1 <-intCor(merged[vv]) plot(merged,xlab="study A",ylab="study B",main="CORRELATION OF CORRELATION",col=3,pch=4) hist(corcor1,xlab="CORRELATION OF CORRELATION") }