spca_randtest {adegenet} | R Documentation |
Monte Carlo test for sPCA
Description
The function spca_randtest
implements Monte-Carlo tests for the
presence of significant spatial structures in a sPCA object. Two tests are
run, for global (positive autocorrelation) and local (negative
autocorrelation) structures, respectively. The test statistics used are the
sum of the absolute values of the corresponding eigenvalues.
Usage
spca_randtest(x, nperm = 499)
Arguments
x |
A |
nperm |
The number of permutations to be used for the test. |
Value
A list with two objects of the class 'randtest' (see
as.randtest
), the first one for 'global' structures
(positivie autocorrelation) and the second for 'local' structures (negative
autocorrelation).
Author(s)
Original code by Valeria Montano adapted by Thibaut Jombart.
Examples
## Not run:
## Load data
data(sim2pop)
## Make spca
spca1 <- spca(sim2pop, type = 1, scannf = FALSE, plot.nb = FALSE)
spca1
plot(spca1)
## run tests (use more permutations in practice, e.g. 999)
tests <- spca_randtest(spca1, nperm = 49)
## check results
tests
plot(tests[[1]]) # global structures
## End(Not run)
[Package adegenet version 2.1.10 Index]