Plotting a phylogeny with the package ggtree (5 min)

Installing the package ggtree (10 min)

• ggtree is an extension of the ggplot2 package, developed specifically for phylogenetic tree visualization
• The author has made available an extensive book with examples
• ggtree is hosted in Bioconductor (not CRAN.
• do length(available()) if you want to know the number of R packages available for installation, both from CRAN and Bioconductor
• “CRAN hosts over 15000 packages and is the official repository for user contributed R packages. Bioconductor provides open source software oriented towards bioinformatics and hosts over 1800 R packages”, from An Introduction to R
• Bioconductor Vs CRAN
• The function install.packages() that we know well only workf for CRAN packages
• To install an R package from Bioconductor, use the function install() from the package BiocManager:
  1. Install BiocManager from CRAN with install.packages("BiocManager")
  2. Then install ggtree from Bioconductor with BiocManager::install("ggtree")

Exercise 1

1. Download this phylogenetic tree of species from the Portal Project Teaching Database, by clicking on the link and saving it to your data-raw folder.
2. Open the file by clicking on its name on the Files tab of RStudio’s Plots pane. It should look like this:
3. Use the main function of the package ggtree (it is also called ggtree()) to visualize portal-tree.tre:

   ggtree(portal_tree)
   

4. What differences can you note between a ggtree() plot and one generated with plot.phylo()?

The phylo class structure (10 min)

• Type the name of the tree you just created and look at the output
• What information is printed to screen?
• Use functions to explore the structure of the objects you just created:
  • class(portal_tree), portal_tree is an object of class "phylo"
  • length(tree), it has length 4
  • names(tree), and it has names
  • Just as with data frames, we can access the named elements of a "phylo" object using the dollar sign $
    • portal_tree$edge, class(portal_tree$edge)
    • portal_tree$Nnode,
    • portal_tree$tip.label
    • portal_tree$node.label
  • str(tree), shows a summary of the elements of the "phylo" object
  • typeof(tree), the "phylo" class is an object of type "list"
• 🎗️ classes and types are data structures that R uses to store/extract information
• a "list" is a data type (or object type), that can hold one or more objects of different types.
• the class "phylo" is a list that combines a matrix, a numeric vector of length one and two character vectors.
• the "phylo" class provides R with all the information it needs to represent a phylogenetic tree

Connecting a phylogeny with data from a table

• Preparation:
  1. Download a data table of the species from the Portal Data base that inlcudes taxonomy
  2. Save it in your data-raw folder.
  3. Read it into R with read.csv(), and assign it to an object called taxonomy.
• To join a tree and a data table, we will use the _join() functions that we used previously to join tables
  • Mini review; example with surveys and species:

    species <- read.csv("../data-raw/species.csv")
    surveys <- read.csv("../data-raw/surveys.csv")
    intersect(colnames(species), colnames(surveys))
    library(dplyr)
    joined_left <- left_join(surveys, species, by = "plot_id")
    joined_inner <- inner_join(surveys, species, by = "plot_id")
    

  • What is the difference between left_join and inner_join?
• To link a tree and a data table, the tree has to be the first argument and the table will be second
• Also, the column that we will be joining by is always "label"
• Make sure there is a “label” column that has some names from your tips in your table

  tree_table <- left_join(portal_tree, taxonomy_matched, by= "label")
  

• What is the structure of the object?
• Attention! doing a full join does not work later down the analysis workflow, we need a left join to drop non matches

  full_join(portal_tree, taxonomy, by = "label")
  

• We can still plot our tree normally with ggtree(tree_table)
• But now we can use aesthetics to color some tip labels by a varianle of choice:

  ggtree(tree, aes(color = taxa, fontface = "italic")) + # it freezes if there are any unmatched or NA labels in data table!!!
  xlim(0, 20) +
  geom_tiplab()
  

Exercise 2: A taxonomy table for small_tree

1. Find the appropriate scientific group labels for each genus in small_tree using this tree as guide.
2. Create a data frame with 3 columns:
   • a "label" column with the names of the tip labels of small_tree. Tip: extract the element "tip.label" from your phylo object to get a vector of tip labels that you can then join to the other vectors to create a data frame.
   • a "taxa" column with the scientific names of the group that each genus belongs to.
   • a "common_name" column with the common names of the group that each genus belongs to. Tip: use the function c() to create the vectors that will be columns "taxa" and "common_name"
3. Join your tree and your table using left_join().
4. Create two different tree plots using taxa and common name to color the tips of the tree.

Exercise 3: Connecting a tree and a table for your final project

1. Connect the data table you chose for your final project to a tree with taxa on your table that you obtained from Open Tree of Life.
2. Plot the tree and color the tips following one variable on the data table.

Homework - Exercise 4: Mapping weight data from surveys CSV table to the portal tree (10 min)

1. Get the average weight and hindfoot length per species.
2. Create a new data frame that contains the taxonomy data plus the averaged data per species that you got on last question.
3. Create two plots with data on the tips, one with the average weight and the other with average hindfoot length. Make sure to also add tip labels, formatted in italics.