varioWeight: Variogram computation with weighting for recurrent genotypes

In ggene: Semivariogram and Exploratory Spatial Analysis of Genetic Data

Description

Compute the omnidirectional variogram for genetic dataset while accounting for recurrent genotypes using a weighting matrix.

Usage

varioWeight(X, weights, return.mat=FALSE,...)

Arguments

X	a ggene object
weights	A matrix of weights. Typically an object produced by genoweight.
return.mat	Logical, if TRUE, the function returns the matrix of weights.
...	arguments to be passed to the function svariog.

.

Details

The function relies on the function variog from package geoR and accepts similar arguments.

Value

A list with 2 items:

svario	an object of class svariog
gamma	a dist object corresponding to the matrix of semi-variance (gene diversity)
weight	the matrix of weights

Author(s)

Jean-Pierre Rossi <[e-mail]>

References

Wagner, H. H., R. Holderegger, S. Werth, F. Gugerli, S. E. Hoebee and C. Scheidegger. 2005. Variogram analysis of the spatial genetic structure of continuous populations using multilocus microsatellite data. Genetics 169, 1739-1752.

See Also

genocount, genoweight, svariog

Examples

data(crypho)

# check sampling scheme
plot(crypho$coord[,1],crypho$coord[,2], asp=1)

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

# compute distance intervals
d <- distlag(dist=crypho$coord, dmin=0,distance.lag=50)

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

# plot the variogram for raw data
plot(wva$svario$u, wva$svario$gamma, col="black", type="b", 
	ylim=range(c(wva$svario$gamma,wva$svario$v)), xlab="distance", ylab="semivariance")

# add the weighted variogram
points(wva$svario$u, wva$svario$v, col="red", type="b", pch=4)

legend("top", legend=c("raw", "weighted"), col=c("black", "red"), lty="solid", pch=c(1,4), bty="n")

## changing distance increment
d <- distlag(dist=crypho$coord, dmin=0,distance.lag=75)
wva <- varioWeight(X=crypho, weights=mat,  uvec=d)

d <- distlag(dist=crypho$coord, dmin=0,distance.lag=50)
wva2 <- varioWeight(X=crypho, weights=mat,  uvec=d)

d <- distlag(dist=crypho$coord, dmin=0,distance.lag=22)
wva3 <- varioWeight(X=crypho, weights=mat,  uvec=d)

plot(wva$svario$u, wva$svario$v, type="b", pch=1, lty=2, xlab="distance (m)",
	ylab="semivariance", ylim=range(c(wva$svario$v,wva2$svario$v,wva3$svario$v)))
points(wva2$svario$u, wva2$svario$v, type="b", pch=17, lty=2)
points(wva3$svario$u, wva3$svario$v, type="b", pch=4, lty=2)
title("weighted variograms for different distance increments")
legend("topleft", legend=c("75m", "50m", "25m"), lty=0, pch=c(1,17,4), bty="n")

## Not run: 

#performs randomization 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
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")


## End(Not run)

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