randsvariog: Performs permutations and computes the variogram statistical...

In ggene: Semivariogram and Exploratory Spatial Analysis of Genetic Data

Description

Randomization of the genetic data with no change of the spatial position of the individuals. The function return the statistical envelope for the specified quantiles.

Usage

randsvariog(var, X, weights=NULL, nsim=99, bounds=NULL, save.sim=FALSE, ...)

Arguments

var	An object of class svariog, typically an output of the function svariog.
X	A ggene object.
weights	A matrix of weights. Typically an object produced by genoweight.
nsim	The number of permutations to be performed.
bounds	2 numerical values indicating the probability for which the quantiles are computed. Defaults set to 0.975 and 0.025.
save.sim	A logical value. If TRUE, the simulated values are to be returned.
...	additional arguments to be passed to the function variog (package geoR).

Details

The function performs permutations of the genetic data while keeping the spatial position of the individuals unchanged. The semivariance is computed for each "randomized" dataset. The upper and lower bounds are derived from the randomized values of the semivariance at each spatial lag.

Value

svario	An object of class "variogram" created with svariog.
env	the quantiles corresponding to the upper and lower bounds;
simul	OPTIONAL. A data.frame where columns contain the values of the semi-variance for each lag for each permutation.

Author(s)

Jean-Pierre Rossi <[e-mail]>

See Also

svariog, genoweight

Examples

data(larix2300)

# check sampling scheme
plot(larix2300$coord[,1],larix2300$coord[,2])

# compute variogram
va <- svariog(X=larix2300, uvec=distlag(dist=larix2300$coord, dmin=0, distance.lag=3),
	plot=FALSE)
plot(va$svario$u, va$svario$v)

## Not run: 
# compute statistical envelope
env <- randsvariog(var=va, X=larix2300, nsim=30, bounds=c(0.025, 0.975), save.sim=FALSE)

# plot results
plot(env$svario$u, env$svario$v, ylim=range(env$env), xlab="distance", ylab="semi-variance")
points(env$svario$u, env$env[,1], type="l")
points(env$svario$u, env$env[,2], type="l")
 
# Repeated genotypes: envelopes for raw and weighted variograms
data(crypho)

#compute the weights
count <- genocount(X=crypho)
mat <- genoweight(X=crypho,genotyp=count$vec)

#performs the randomizations on raw variogram
va <- svariog(X=crypho, plot=FALSE)
env <- randsvariog(var=va, X=crypho, nsim=30, bounds=NULL, save.sim=FALSE)

#compute the weighted variogram
wva <- varioWeight(X=crypho, weights=mat)

#performs the randomizations on weighted variogram
env2 <- randsvariog(var=wva, X=crypho, nsim=30, bounds=NULL, save.sim=FALSE, weights=mat)

# Plot results

# plot results
xx <- c(wva$svario$u, rev(wva$svario$u))
yy <- c(env$env[,1], rev(env$env[,2]))
plot(xx, yy, type = "n", xlab = "distance", ylab = "semivariance",
 ylim=range(c(env$env[,1], env$env[,2], env2$env[,1], env2$env[,2])))
polygon(xx, yy, col = "lightgrey", border = "black")
xx <- c(wva$svario$u, rev(wva$svario$u))
yy <- c(env2$env[,1], env2$env[,2])
points(xx, yy, type = "l")
polygon(xx, yy, col = "lightblue", border = "blue")

points(wva$svario$u, wva$svario$v, col="blue", typ="b")
points(wva$svario$u, wva$svario$gamma, col="black", type="b", lty="solid", bty="n")

legend("top", legend=c("raw", "weighted"), col=c("black", "blue"), lty="solid", bty="n")

## End(Not run) 
      

ggene documentation built on May 2, 2019, 5:54 p.m.