| permutest {CORREP} | R Documentation |
This function calculates p-values of the multivariate correlation estimator by enumerating all permutations. We recommend using Likehood Ratio Test implemented in function cor.LRtest1 if your data has moderate to large sample size (>5) The procedure is same as those permutation tests for Pearson correlation coefficient or other parameters. Since the approximation of null distribution requires enumerating all permutations. The computational burden increases in $n^2$.
permutest(x, y=NULL, m, G)
x |
data matrix, column represents samples (conditions), and row represents variables (genes), see example below for format information |
y |
optional, used when x and y are vectors |
m |
number of replicates |
G |
number of genes |
See manuscript.
PV |
P-values of permutation tests |
Dongxiao Zhu and Youjuan Li
Zhu, D and Li Y. 2007. Multivariate Correlation Estimator for Inferring Functional Relationships from Replicated 'OMICS' data. Submitted.
cor.LRtest1, cor.LRtest2, cor.test
library("CORREP")
library("e1071")
d0 <- NULL
## sample size is set to 5, it takes about a min to finish
for(l in 1:5)
d0 <- rbind(d0, rnorm(100))
## data must have row variance of 1
d0.std <- apply(d0, 2, function(x) x/sd(x))
M <- cor.balance(t(d0.std), m = 4, G= 25)
M.pv <- permutest(t(d0.std), m = 4, G= 25)