mgroup: Mantel test for groups

Description

Mantel test across one or more group contrasts.

Usage

mgroup(edist, groups, nperm = 1000, mrank = FALSE)

Value

nclust: Number of groups tested.
mantelr: Mantel coefficient.
pval1: one-tailed p-value (null hypothesis: r <= 0).

Arguments

edist: a dist object or lower triangular distance vector.
groups: a vector of group memberships (numeric, character, or factor), or a matrix or data frame with columns describing multiple sets of groups.
nperm: number of permutations to use. If set to 0, the permutation test will be omitted.
mrank: if this is set to FALSE (the default option), Pearson correlations will be used. If set to TRUE, the Spearman correlation (correlation ranked distances) will be used.

Author

Sarah Goslee

Details

mgroup returns the Mantel correlations for group contrast matrices computed from cluster groups across a range of clustering levels.

References

Legendre, P. and M. Fortin. 1989. Spatial pattern and ecological analysis. Vegetatio 80:107-138.

Examples

Run this code


# Using a model matrix to test group membership

data(iris)
iris.d <- dist(iris[,1:4])
mgroup(iris.d, iris[,5])

# clustering-based example

data(graze)
graze.d <- dist(graze[, -c(1:2)])
graze.hclust <- hclust(graze.d)

clust.groups <- data.frame(
	k2 = cutree(graze.hclust, k = 2),
	k4 = cutree(graze.hclust, k = 4),
	k6 = cutree(graze.hclust, k = 6),
	k8 = cutree(graze.hclust, k = 8))

mgroup(graze.d, clust.groups, nperm=1000)

Run the code above in your browser using DataLab

Data engineering and BI courses are free!