tanglegram: Tanglegram plot

View source: R/tanglegram.R

tanglegramR Documentation

Tanglegram plot

Description

Plots a tanglegram plot of a side by side trees.

Usage

tanglegram(dend1, ...)

## Default S3 method:
tanglegram(dend1, ...)

## S3 method for class 'hclust'
tanglegram(dend1, ...)

## S3 method for class 'phylo'
tanglegram(dend1, ...)

## S3 method for class 'dendlist'
tanglegram(
  dend1,
  which = c(1L, 2L),
  main_left,
  main_right,
  just_one = TRUE,
  ...
)

## S3 method for class 'dendrogram'
tanglegram(
  dend1,
  dend2,
  sort = FALSE,
  color_lines,
  lwd = 3.5,
  edge.lwd = NULL,
  columns_width = c(5, 3, 5),
  margin_top = 3,
  margin_bottom = 2.5,
  margin_inner = 3,
  margin_outer = 0.5,
  left_dendo_mar = c(margin_bottom, margin_outer, margin_top, margin_inner),
  right_dendo_mar = c(margin_bottom, margin_inner, margin_top, margin_outer),
  intersecting = TRUE,
  dLeaf = NULL,
  dLeaf_left = dLeaf,
  dLeaf_right = dLeaf,
  axes = TRUE,
  type = "r",
  lab.cex = NULL,
  remove_nodePar = FALSE,
  main = "",
  main_left = "",
  main_right = "",
  sub = "",
  k_labels = NULL,
  k_branches = NULL,
  rank_branches = FALSE,
  hang = FALSE,
  match_order_by_labels = TRUE,
  cex_main = 2,
  cex_main_left = cex_main,
  cex_main_right = cex_main,
  cex_sub = cex_main,
  highlight_distinct_edges = TRUE,
  common_subtrees_color_lines = TRUE,
  common_subtrees_color_lines_default_single_leaf_color = "grey",
  common_subtrees_color_branches = FALSE,
  highlight_branches_col = FALSE,
  highlight_branches_lwd = TRUE,
  faster = FALSE,
  just_one = TRUE,
  ...
)

dendbackback(
  dend1,
  dend2,
  sort = FALSE,
  color_lines,
  lwd = 3.5,
  edge.lwd = NULL,
  columns_width = c(5, 3, 5),
  margin_top = 3,
  margin_bottom = 2.5,
  margin_inner = 3,
  margin_outer = 0.5,
  left_dendo_mar = c(margin_bottom, margin_outer, margin_top, margin_inner),
  right_dendo_mar = c(margin_bottom, margin_inner, margin_top, margin_outer),
  intersecting = TRUE,
  dLeaf = NULL,
  dLeaf_left = dLeaf,
  dLeaf_right = dLeaf,
  axes = TRUE,
  type = "r",
  lab.cex = NULL,
  remove_nodePar = FALSE,
  main = "",
  main_left = "",
  main_right = "",
  sub = "",
  k_labels = NULL,
  k_branches = NULL,
  rank_branches = FALSE,
  hang = FALSE,
  match_order_by_labels = TRUE,
  cex_main = 2,
  cex_main_left = cex_main,
  cex_main_right = cex_main,
  cex_sub = cex_main,
  highlight_distinct_edges = TRUE,
  common_subtrees_color_lines = TRUE,
  common_subtrees_color_lines_default_single_leaf_color = "grey",
  common_subtrees_color_branches = FALSE,
  highlight_branches_col = FALSE,
  highlight_branches_lwd = TRUE,
  faster = FALSE,
  just_one = TRUE,
  ...
)

Arguments

dend1

tree object (dendrogram/dendlist/hclust/phylo), plotted on the left

...

not used.

which

an integer vector of length 2, indicating which of the trees in the dendlist object should be plotted

main_left

Character. Title of the left dendrogram.

main_right

Character. Title of the right dendrogram.

just_one

logical (TRUE). If FALSE, it means at least two tanglegrams will be plotted on the same page and so layout is not passed. See: https://stackoverflow.com/q/39784746/4137985

dend2

tree object (dendrogram/hclust/phylo), plotted on the right

sort

logical (FALSE). Should the dendrogram's labels be "sorted"? (might give a better tree in some cases).

color_lines

a vector of colors for the lines connected the labels. If the colors are shorter than the number of labels, they are recycled (and a warning is issued). The colors in the vector are applied on the lines from the bottom up.

lwd

width of the lines connecting the labels. (default is 3.5)

edge.lwd

width of the dendrograms lines. Default is NULL. If set, then it switches 'highlight_branches_lwd' to FALSE. If you want thicker lines which reflect the height, please use highlight_branches_lwd on the dendrograms/dendlist.

columns_width

a vector with three elements, giving the relative sizes of the the three plots (left dendrogram, connecting lines, right dendrogram). This is passed to layout if parameter just_one is TRUE. The default is: c(5,3,5)

margin_top

the number of lines of margin to be specified on the top of the plots.

margin_bottom

the number of lines of margin to be specified on the bottom of the plots.

margin_inner

margin_bottom the number of lines of margin to be specified on the inner distence between the dendrograms and the connecting lines.

margin_outer

margin_bottom the number of lines of margin to be specified on the outer distence between the dendrograms and the connecting lines.

left_dendo_mar

mar parameters of the left dendrgoram.

right_dendo_mar

mar parameters of the right dendrgoram.

intersecting

logical (TRUE). Should the leaves of the two dendrograms be pruned so that the two trees will have the same labels?

dLeaf

a number specifying the distance in user coordinates between the tip of a leaf and its label. If NULL, as per default, 3/4 of a letter width or height is used.

Notice that if we are comparing two dendrograms with different heights, manually changing dLeaf will affect both trees differently. In such a case, it is recommanded to manually change dLeaf_left and dLeaf_right. This can be especially important when changing the lab.cex of the dendrogram's labels. Alternatively, one could manually set the xlim parameter for both trees, which will force the proportion of distances of the labels from the trees to remain the same.

dLeaf_left

dLeaf of the left dendrogram, by default it is equal to dLeaf (often negative).

dLeaf_right

dLeaf of the right dendrogram, by default it is equal to minus dLeaf (often positive).

axes

logical (TRUE). Should plot axes be plotted?

type

type of plot ("t"/"r" = triangle or rectangle)

lab.cex

numeric scalar, influanicing the cex size of the labels.

remove_nodePar

logical (FALSE). Should the nodePar of the leaves be removed? (useful when the trees' leaves has too many parameters on them)

main

Character. Title above the connecting lines.

sub

Character. Title below the connecting lines.

k_labels

integer. Number of groups by which to color the leaves.

k_branches

integer. Number of groups by which to color the branches.

rank_branches

logical (FALSE). Should the branches heights be adjusted? (setting this to TRUE - can make it easier for comparing topological differences)

hang

logical (FALSE). Should we hang the leaves of the trees?

match_order_by_labels

logical (TRUE). Should the leaves value order be matched between the two trees based on labels? This is a MUST in order to have the lines connect the correct labels. Set this to FALSE if you want to make the plotting a bit faster, and only after you are sure the labels and orders are correctly aligned.

cex_main

A numerical value giving the amount by which plotting title should be magnified relative to the default.

cex_main_left

see cex_main.

cex_main_right

see cex_main.

cex_sub

see cex_main.

highlight_distinct_edges

logical (default is TRUE). If to highlight distinct edges in each tree (by changing their line types to 2). (notice that this can be slow on large trees)

This parameter will automatically be turned off if the tree already comes with a "lty" edgePar (this is checked using has_edgePar). A "lty" can be removed by using set("clear_branches"), by removing all of the edgePar parameters of the dendrogram.

common_subtrees_color_lines

logical (default is TRUE). color the connecting line based on the common subtrees of both dends. This only works if (notice that this can be slow on large trees)

common_subtrees_color_lines_default_single_leaf_color

When representing edges between common subtrees (i.e. common_subtrees_color_branches = TRUE), this parameter sets the color of edges for subtrees that are NOT common. Default is "grey"

common_subtrees_color_branches

logical (default is FALSE). Color the branches of both dends based on the common subtrees. (notice that this can be slow on large trees) This is FALSE by default since it will override the colors of the existing tree.

highlight_branches_col

logical (default is FALSE). Should highlight_branches_col be used on the dendrograms.

This parameter will automatically be turned off if the tree already comes with a "col" edgePar (this is checked using has_edgePar). A "lty" can be removed by using set("clear_branches"), by removing all of the edgePar parameters of the dendrogram.

highlight_branches_lwd

logical (default is TRUE). Should highlight_branches_lwd be used on the dendrograms.

This parameter will automatically be turned off if the tree already comes with a "lwd" edgePar (this is checked using has_edgePar). A "lty" can be removed by using set("clear_branches"), by removing all of the edgePar parameters of the dendrogram.

faster

logical (FALSE). If TRUE, it overrides some other parameters to have them turned off so that the plotting will go a tiny bit faster.

Details

Notice that tanglegram does not "resize" well. In case you are resizing your window you would need to re-run the function.

Value

An invisible dendlist, with two trees after being modified during the creation of the tanglegram.

Author(s)

Tal Galili, Johan Renaudie

Source

The function is based on code from Johan Renaudie (plannapus), after major revisions. See: https://stackoverflow.com/questions/12456768/duelling-dendrograms-in-r-placing-dendrograms-back-to-back-in-r

As far as I could tell, this code was originally inspired by Dylan Beaudette function dueling.dendrograms from the sharpshootR package: https://CRAN.R-project.org/package=sharpshootR tanglegram

See Also

remove_leaves_nodePar, plot_horiz.dendrogram, rank_branches, hang.dendrogram

Examples

## Not run: 
set.seed(23235)
ss <- sample(1:150, 10)
dend1 <- iris[ss, -5] %>%
  dist() %>%
  hclust("com") %>%
  as.dendrogram()
dend2 <- iris[ss, -5] %>%
  dist() %>%
  hclust("sin") %>%
  as.dendrogram()
dend12 <- dendlist(dend1, dend2)

dend12 %>% tanglegram()

tanglegram(dend1, dend2)
tanglegram(dend1, dend2, sort = TRUE)
tanglegram(dend1, dend2, remove_nodePar = TRUE)
tanglegram(dend1, dend2, k_labels = 6, k_branches = 4)

tanglegram(dend1, dend2,
  lab.cex = 2, edge.lwd = 3,
  margin_inner = 5, type = "t", center = TRUE
)


## works nicely:
tanglegram(dend1, dend2,
  lab.cex = 2, edge.lwd = 3,
  margin_inner = 3.5, type = "t", center = TRUE,
  dLeaf = -0.1, xlim = c(7, 0),
  k_branches = 3
)


# using rank_branches can make the comparison even easier
tanglegram(rank_branches(dend1), rank_branches(dend2),
  lab.cex = 2, edge.lwd = 3,
  margin_inner = 3.5, type = "t", center = TRUE,
  dLeaf = -0.1, xlim = c(5.1, 0), columns_width = c(5, 1, 5),
  k_branches = 3
)



########
## Nice example of some colored trees

# see the coloring of common sub trees:
set.seed(23235)
ss <- sample(1:150, 10)
dend1 <- iris[ss, -5] %>%
  dist() %>%
  hclust("com") %>%
  as.dendrogram()
dend2 <- iris[ss, -5] %>%
  dist() %>%
  hclust("sin") %>%
  as.dendrogram()
dend12 <- dendlist(dend1, dend2)
# dend12 %>% untangle %>% tanglegram
dend12 %>% tanglegram(common_subtrees_color_branches = TRUE)


set.seed(22133513)
ss <- sample(1:150, 10)
dend1 <- iris[ss, -5] %>%
  dist() %>%
  hclust("com") %>%
  as.dendrogram()
dend2 <- iris[ss, -5] %>%
  dist() %>%
  hclust("sin") %>%
  as.dendrogram()
dend12 <- dendlist(dend1, dend2)
# dend12 %>% untangle %>% tanglegram
dend12 %>% tanglegram(common_subtrees_color_branches = TRUE)
dend12 %>% tanglegram()

## End(Not run)

talgalili/dendextend documentation built on Jan. 27, 2024, 7:43 p.m.