Plot a hierarchical clustering bioregionalization

Plots the output of a clustering function. For hierarchical clustering (hclu_hierarclust(), hclu_diana() and cut_tree()), the tree is drawn with coloured bars under it showing the cluster of each site. Plots are not available for the other clustering algorithms. Arguments starting with tree_ apply to trees only.

Usage

# S3 method for class 'bioregion.clusters'
plot(
  x,
  tree_branch_partition = NULL,
  tree_labels = TRUE,
  tree_flip = FALSE,
  tree_horizontal = FALSE,
  ...
)

Arguments

x: A bioregion.clusters object.
tree_branch_partition: Trees only. The partition whose clusters give their colours to the branches of the tree: either a partition number (1 for the first partition, as in map_bioregions()) or the name of a partition (e.g., "K_8"). By default (NULL), the finest partition (i.e. with the most clusters) is used.
tree_labels: Trees only. A boolean indicating whether the site names should be written next to the bars (default TRUE).
tree_flip: Trees only. A boolean indicating whether the order of the sites should be reversed, i.e., the tree drawn as in a mirror (default FALSE). The tree remains the same: only the two branches of each node are swapped.
tree_horizontal: Trees only. A boolean indicating whether the tree should be drawn horizontally, with its root on the left and the sites on the right, which makes site names easier to read (default FALSE).
...: Trees only. Further arguments passed to graphics::text() to customize the site names (e.g., cex to change their size).

Details

Tree plots (currently, only hclu_ functions)

The tree is drawn with one coloured bar for each partition (i.e., each cut of the tree), from coarsest to finest cuts. Each bar is labelled with the name of its partition. In each bar, the colour next to a site shows the cluster of that site. Sites that were not assigned to any cluster (NA) are left blank.

The branches that contain sites of a single cluster of tree_branch_partition are drawn in the colour of that cluster. The other branches, which join several clusters, are black. When the tree was cut at given heights (or heights found for given numbers of clusters), dashed lines show these heights.

With the dynamic tree cut (cut_tree(dynamic_tree_cut = TRUE)), the clusters are not obtained by cutting the whole tree at a single height, and a cluster is not always a whole branch of the tree: it can be made of neighbouring sites hanging from different branches. With dynamic_method = "hybrid", a cluster can even appear in several places of the tree. The bars show these clusters as they are.

The colours are those stored in x by bioregion_colors(), so that the tree and the maps made with map_bioregions() share the same colours. If x has no colours, or if its colours do not match its partitions (e.g., after cutting the tree again), colours are assigned with bioregion_colors() and its default parameters.

The site names are written in the bottom margin of the plot, or in its right margin if tree_horizontal = TRUE. Long site names may not fit in it: either reduce their size (e.g., cex = 0.5), enlarge the margin with graphics::par() before plotting (e.g., par(mar = c(8, 4, 1, 1)), or par(mar = c(4, 1, 4, 8)) for a horizontal tree), or remove them with tree_labels = FALSE.

Author

Boris Leroy (leroy.boris@gmail.com)
Maxime Lenormand (maxime.lenormand@inrae.fr)
Pierre Denelle (pierre.denelle@gmail.com)

Examples

data(fishmat)
dissim <- dissimilarity(fishmat, metric = "Simpson")

# One partition
tree1 <- hclu_hierarclust(dissim, n_clust = 8,
                          optimal_tree_method = "best", verbose = FALSE)
plot(tree1, cex = 0.3)


# Several partitions, branches coloured with the partition with the fewest
# clusters (column 2 of tree2$clusters)
tree2 <- cut_tree(tree1, n_clust = c(4, 8, 16), verbose = FALSE)
plot(tree2, tree_branch_partition = 2, tree_labels = FALSE)


# Same tree, mirrored
plot(tree2, tree_flip = TRUE, tree_labels = FALSE)


# Dynamic tree cut
tree3 <- cut_tree(tree1, dynamic_tree_cut = TRUE, verbose = FALSE)
plot(tree3, tree_labels = FALSE)



# Same colours as on maps
tree3 <- bioregion_colors(tree2,
                          palette = "Antique")
plot(tree3, tree_labels = FALSE)

if (FALSE) { # \dontrun{
data(fishsf)
map_bioregions(tree3, fishsf)
} # }

# Horizontal tree, with room for the site names in the right margin
small_mat <- fishmat[1:30, ]
small_mat <- small_mat[, colSums(small_mat) > 0]
tree4 <- hclu_hierarclust(dissimilarity(small_mat, metric = "Simpson"),
                          n_clust = c(3, 6), optimal_tree_method = "best",
                          verbose = FALSE)
old_par <- par(mar = c(4, 1, 4, 7))
plot(tree4, tree_horizontal = TRUE, cex = 0.7)

par(old_par)

Usage

Arguments

Details

See also

Author

Examples