EcoGenetics (version 1.2.1-3)

eco.post.geneland: Log posterior probability plot for Geneland repetitions with fixed K

Description

Log posterior probability plot for Geneland repetitions with fixed K

Usage

eco.post.geneland(niter, burnin)

Arguments

niter

Number of mcmc iterations per repetition.

burnin

Number of mcmc to burn-in.

Details

This program returns, for a series of Geneland repetitions with fixed K, and a specified burn-in value, a plot of the log posterior probability vs the repetition number. This allows to choose the best run. The working directory will be set to the folder containing the results created by Geneland. The program expects each subfolder (run) to have a number as name, that indicates the corresponding number of run. (1, 2, etc., see the example).

Examples

# NOT RUN {
require("Geneland")
data(eco.test)

# We create a folder in the working directory for the results and 
# save the data frames of the object "eco" in the format required
# by Geneland:

path.1 <- getwd()
path <- paste(path.1,"/test/", sep="")
dir.create(path) 
setwd(path)
ecogen2geneland(eco, ploidy = 2)

# Auxiliar function for running some repetitions with fixed K = 4.
# Each repetition is saved in the folder "test":
simul <- function(i) {
 path <- getwd()
path <- paste(path,"/",i, sep = "")
dir.create(path) 
MCMC(coordinates = read.table("XY.txt"), 
		 geno.dip.codom = read.table("G.txt"), 
		 varnpop = TRUE, npopmin = 4, npopmax = 4, spatial = TRUE, 
		 freq.model = "Correlated", nit = 500, thinning = 10,
		 path.mcmc = path)
}

# 5 repetitions with K = 4 
lapply(1:5, simul)

# Check that in the folder "test" are the simulated result.
# Your results must have that appearance.

# Plot of the repetition order number vs the corresponding
# posterior probability, with a burn-in of 10 mcmc:
eco.post.geneland(5, 10)
# }
# NOT RUN {
# }