TreeDist (version 2.9.2)

Islands: Find islands from distance matrix

Description

Islands() assigns a set of objects to islands, such that all elements within an island can form a connected graph in which each edge is no longer than threshold distance units Silva2021TreeDist.

Usage

Islands(D, threshold, dense = TRUE, smallest = 0)

Value

Islands() returns a vector listing the island to which each element is assigned.

Arguments

D

Square matrix or dist object containing Euclidean distances between data points.

threshold

Elements greater than threshold distance units will not be assigned to the same island.

dense

Logical; if FALSE, each island will be named according to the index of its lowest-indexed member; if TRUE, each island will be numbered sequentially from 1, ordered by the index of the lowest-indexed member.

smallest

Integer; Islands comprising no more than smallest elements will be assigned to island NA.

Author

Martin R. Smith ()

References

See Also

Other tree space functions: MSTSegments(), MapTrees(), MappingQuality(), SpectralEigens(), cluster-statistics, median.multiPhylo()

Examples

library("TreeTools", quietly = TRUE)
# Generate a set of trees
trees <- as.phylo(as.TreeNumber(BalancedTree(16)) + c(-(40:20), 70:105), 16)

# Calculate distances between trees
distances <- ClusteringInfoDist(trees)
summary(distances)

# Assign trees to islands
isle <- Islands(distances, quantile(distances, 0.1))
table(isle)

# Indicate island membership on 2D mapping of tree distances
mapping <- cmdscale(distances, 2)
plot(mapping, col = isle + 1,
     asp = 1, # Preserve aspect ratio - do not distort distances
     ann = FALSE, axes = FALSE, # Don't label axes: dimensions are meaningless)
     pch = 16 # Plotting character: Filled circle
)

# Compare strict consensus with island consensus trees
oPar <- par(mfrow = c(2, 2), mai = rep(0.1, 4))
plot(Consensus(trees), main = "Strict")
plot(Consensus(trees[isle == 1]), edge.col = 2, main = "Island 1")
plot(Consensus(trees[isle == 2]), edge.col = 3, main = "Island 2")
plot(Consensus(trees[isle == 3]), edge.col = 4, main = "Island 3")

# Restore graphical parameters
par(oPar)