gstat.randtest {adegenet}R Documentation

Goudet's G-statistic Monte Carlo test for genind object

Description

The function gstat.randtest implements Goudet's G-statistic Monte Carlo test (g.stats.glob, package hierfstat) for genind object.
The output is an object of the class randtest (package ade4) from a genind object.

This procedure tests for genetic structuring of individuals using 3 different schemes (see details).

Usage

gstat.randtest(x,pop=NULL, method=c("global","within","between"),
sup.pop=NULL, sub.pop=NULL, nsim=499)

Arguments

x an object of class genind.
pop a factor giving the 'population' of each individual. If NULL, pop is seeked from x@pop. Note that the term population refers in fact to any grouping of individuals'.
method a character (if a vector, only first argument is kept) giving the method to be applied: 'global', 'within' or 'between' (see details).
sup.pop a factor indicating any grouping of individuals at a larger scale than 'pop'. Used in 'within' method.
sub.pop a factor indicating any grouping of individuals at a finer scale than 'pop'. Used in 'between' method.
nsim number of simulations to be used for the randtest.

Details

This G-statistic Monte Carlo procedure tests for population structuring at different levels. This is determined by the argument 'method':

- "global": tests for genetic structuring given 'pop'.
- "within": tests for genetic structuring within 'pop' inside each 'sup.pop' group (i.e., keeping sup.pop effect constant).
- "between": tests for genetic structuring between 'pop' keeping individuals in their 'sub.pop' groups (i.e., keeping sub.pop effect constant).

Value

Returns an object of the class randtest (package ade4).

Author(s)

Thibaut Jombart jombart@biomserv.univ-lyon1.fr

See Also

g.stats.glob, fstat, test.g,test.within, test.between,as.randtest, genind2hierfstat

Examples

if(require(hierfstat)){
# here the example of g.stats.glob is taken using gstat.randtest
data(gtrunchier)
x <- df2genind(X=gtrunchier[,-c(1,2)],pop=gtrunchier$Patch)

# test in hierfstat
gtr.test<- g.stats.glob(gtrunchier[,-1])
gtr.test

# randtest version
x.gtest <- gstat.randtest(x,nsim=99)
x.gtest
plot(x.gtest)

# pop within sup.pop test
gstat.randtest(x,nsim=99,method="within",sup.pop=gtrunchier$Locality)

# pop test with sub.pop kept constant
gstat.randtest(x,nsim=99,pop=gtrunchier$Locality,method="between",sub.pop=gtrunchier$Patch)
}

[Package adegenet version 1.2-2 Index]