Files
ROBITools/man/m.univariate.test.Rd

48 lines
1.6 KiB
R

% Generated by roxygen2: do not edit by hand
% Please edit documentation in R/mstat.R
\name{m.univariate.test}
\alias{m.univariate.test}
\title{Test the significance of the M statistics by Monte-Carlo}
\usage{
m.univariate.test(w, groups, resampling = 100, alternative = "two.sided")
}
\arguments{
\item{w}{the weigth matrix indicating the presence probability of each motu
in each samples. Each line corresponds to a sample and each column
to a MOTU. \code{rownames} of the \code{w} matrix must be the sample
names.}
\item{groups}{the list of considered groups as computed by the \code{\link{dist.center.group}}
function}
\item{resampling}{the number of simulation to establish the null distribution}
\item{alternative}{a character value in \code{c('two.sided','less','greater')}
- two.sided : the m stat is check against the two side of the empirical
M distribution
- less : test if the M stat is lesser than the M observed in the the empirical
M distribution (exlusion hypothesis)
- greater : test if the M stat is greater than the M observed in the the empirical
M distribution (aggregation hypothesis)}
}
\value{
a vector of p.value with an attribute \code{m.stat} containing the actual M stat
for each MOTUs
}
\description{
Computes computes the p.value the M statistics asociated to a MOTU
by shuffling MOTUs among location.
}
\examples{
data(termes)
termes.ok = termes[,colSums(termes$reads)>0]
pos = expand.grid(1:3 * 10,1:7 * 10)
labels = rownames(termes.ok)
d = dist.grid(pos[,1],pos[2],labels)
groups = dist.center.group(d,20)
w = m.weight(termes.ok)
pval = m.univariate.test(w,groups)
}