How to use rotl?

rotl provides an interface to the Open Tree of Life (OTL) API and allows users to query the API, retrieve parts of the Tree of Life and integrate these parts with other R packages.

The OTL API provides services to access:

In rotl, each of these services correspond to functions with different prefixes:

| Service | rotl prefix | |---------------|---------------| | Tree of Life | tol_ | | TNRS | tnrs_ | | Taxonomy | taxonomy_ | | Studies | studies_ |

rotl also provides a few other functions and methods that can be used to extract relevant information from the objects returned by these functions.

Demonstration of a basic workflow

The most common use for rotl is probably to start from a list of species and get the relevant parts of the tree for these species. This is a two step process:

  1. the species names need to be matched to their ott_id (the Open Tree Taxonomy identifiers) using the Taxonomic name resolution services (TNRS)
  2. these ott_id will then be used to retrieve the relevant parts of the Tree of Life.

Step 1: Matching taxonomy to the ott_id

Let's start by doing a search on a diverse group of taxa: a tree frog (genus Hyla), a fish (genus Salmo), a sea urchin (genus Diadema), and a nautilus (genus Nautilus).

library(rotl)
taxa <- c("Hyla", "Salmo", "Diadema", "Nautilus")
resolved_names <- tnrs_match_names(taxa)

It's always a good idea to check that the resolved names match what you intended:

r knitr::kable(resolved_names)

The column unique_name sometimes indicates the higher taxonomic level associated with the name. The column number_matches indicates the number of ott_id that corresponds to a given name. In this example, our search on Diadema returns 2 matches, and the one returned by default is indeed the sea urchin that we want for our query. The argument context_name allows you to limit the taxonomic scope of your search. Diadema is also the genus name of a fungus. To ensure that our search is limited to animal names, we could do:

resolved_names <- tnrs_match_names(taxa, context_name = "Animals")

If you are trying to build a tree with deeply divergent taxa that the argument context_name cannot fix, see "How to change the ott ids assigned to my taxa?" in the FAQ below.

Step 2: Getting the tree corresponding to our taxa

Now that we have the correct ott_id for our taxa, we can ask for the tree using the tol_induced_subtree() function. By default, the object returned by tol_induced_subtree is a phylo object (from the ape package), so we can plot it directly.

my_tree <- tol_induced_subtree(ott_ids = resolved_names$ott_id)
plot(my_tree, no.margin = TRUE)

FAQ

How to change the ott ids assigned to my taxa?

If you realize that tnrs_match_names assigns the incorrect taxonomic group to your name (e.g., because of synonymy) and changing the context_name does not help, you can use the function inspect. This function takes the object resulting from tnrs_match_names(), and either the row number, the taxon name (you used in your search in lowercase), or the ott_id returned by the initial query.

To illustrate this, let's re-use the previous query but this time pretending that we are interested in the fungus Diadema and not the sea urchin:

taxa <- c("Hyla", "Salmo", "Diadema", "Nautilus")
resolved_names <- tnrs_match_names(taxa)
resolved_names
inspect(resolved_names, taxon_name = "diadema")

In our case, we want the second row in this data frame to replace the information that initially matched for Diadema. We can now use the update() function, to change to the correct taxa (the fungus not the sea urchin):

resolved_names <- update(resolved_names,
  taxon_name = "diadema",
  new_row_number = 2
)

## we could also have used the ott_id to replace this taxon:
## resolved_names <- update(resolved_names, taxon_name = "diadema",
##                          new_ott_id = 4930522)

And now our resolved_names data frame includes the taxon we want:

r knitr::kable(resolved_names)

How do I know that the taxa I'm asking for is the correct one?

The function taxonomy_taxon_info() takes ott_ids as arguments and returns taxonomic information about the taxa. This output can be passed to some helpers functions to extract the relevant information. Let's illustrate this with our Diadema example

diadema_info <- taxonomy_taxon_info(631176)
tax_rank(diadema_info)
synonyms(diadema_info)
tax_name(diadema_info)

In some cases, it might also be useful to investigate the taxonomic tree descending from an ott_id to check that it's the correct taxon and to determine the species included in the Open Tree Taxonomy:

diadema_tax_tree <- taxonomy_subtree(631176)
diadema_tax_tree

By default, this function return all taxa (including self, and internal) descending from this ott_id but it also possible to return phylo object.

How do I get the tree for a particular taxonomic group?

If you are looking to get the tree for a particular taxonomic group, you need to first identify it by its node id or ott id, and then use the tol_subtree() function:

mono_id <- tnrs_match_names("Monotremata")
mono_tree <- tol_subtree(ott_id = ott_id(mono_id))
plot(mono_tree)

How do I find trees from studies focused on my favourite taxa?

The function studies_find_trees() allows the user to search for studies matching a specific criteria. The function studies_properties() returns the list of properties that can be used in the search.

furry_studies <- studies_find_studies(property = "ot:focalCladeOTTTaxonName", value = "Mammalia")
furry_ids <- furry_studies$study_ids

Now that we know the study_id, we can ask for the meta data information associated with this study:

furry_meta <- get_study_meta("pg_2550")
get_publication(furry_meta) ## The citation for the source of the study
get_tree_ids(furry_meta) ## This study has 10 trees associated with it
candidate_for_synth(furry_meta) ## None of these trees are yet included in the OTL

Using get_study("pg_2550") would returns a multiPhylo object (default) with all the trees associated with this particular study, while get_study_tree("pg_2550", "tree5513") would return one of these trees.

The tree returned by the API has duplicated tip labels, how can I work around it?

You may encounter the following error message:

Error in rncl(file = file, ...) : Taxon number 39 (coded by the token Pratia
angulata) has already been encountered in this tree. Duplication of taxa in a
tree is prohibited.

This message occurs as duplicate labels are not allowed in the NEXUS format and it is stricly enforced by the part of the code used by rotl to import the trees in memory.

If you use a version of rotl more recent than 0.4.1, this should not happen by default for the function get_study_tree. If it happens with another function, please let us know.

The easiest way to work around this is to save the tree in a file, and use APE to read it in memory:

get_study_tree(
  study_id = "pg_710", tree_id = "tree1277",
  tip_label = "ott_taxon_name", file = "/tmp/tree.tre",
  file_format = "newick"
)
tr <- ape::read.tree(file = "/tmp/tree.tre")

How do I get the higher taxonomy for a given taxa?

If you encounter a taxon name you are not familiar with, it might be useful to obtain its higher taxonomy to see where it fits in the tree of life. We can combine several taxonomy methods to extract this information easily.

giant_squid <- tnrs_match_names("Architeuthis")
tax_lineage(taxonomy_taxon_info(ott_id(giant_squid), include_lineage = TRUE))

Why are OTT IDs discovered with rotl missing from an induced subtree?

Some taxonomic names that can be retrieved through the taxonomic name resolution service are not part of the Open Tree's synthesis tree. These are usually traditional higher-level taxa that have been found to be paraphyletic.

For instance, if you wanted to fetch a tree relating the three birds that go into a Turkducken as well as the pork used for stuffing, you might search for the turkey, duck, chicken, and pork genera:

turducken <- c("Meleagris", "Anas", "Gallus", "Sus")
taxa <- tnrs_match_names(turducken, context = "Animals")
taxa

We have the OTT ids for each genus, however, if we tried to get the induced subtree from these results, we would get an error:

tr <- tol_induced_subtree(ott_id(taxa))

As the error message suggests, some of the taxa are not found in the synthetic tree. This occurs for 2 main reasons: either the taxa is invalid, or it is part of a group that is not monophyletic in the synthetic tree. There are two ways to get around this issue: (1) removing the taxa that are not part of the Open Tree; (2) using the complete species name.

Removing the taxa missing from the synthetic tree

To help with this situation, rotl provides a way to identify the OTT ids that are not part of the synthetic tree. The function is_in_tree() takes the output of the ott_id() function and returns a vector of logical indicating whether the taxa are part of the synthetic tree. We can then use to only keep the taxa that appear in the synthetic tree:

in_tree <- is_in_tree(ott_id(taxa))
in_tree
tr <- tol_induced_subtree(ott_id(taxa)[in_tree])

Using the full taxonomic names

The best way to avoid these problems is to specify complete species names (species being the lowest level of classification in the Open Tree taxonomy they are guaranteed to be monophyletic):

turducken_spp <- c("Meleagris gallopavo", "Anas platyrhynchos", "Gallus gallus", "Sus scrofa")
taxa <- tnrs_match_names(turducken_spp, context = "Animals")
tr <- tol_induced_subtree(ott_id(taxa))
plot(tr)


Try the rotl package in your browser

Any scripts or data that you put into this service are public.

rotl documentation built on Oct. 23, 2020, 8:22 p.m.