rOpenSci | rotl paper published

We are excited to announce a paper describing rotl, our package for the Open Tree of Life data, has been published. The full citation is:

Michonneau, F., Brown, J. W. and Winter, D. J. (2016), rotl: an R package to interact with the Open Tree of Life data. Methods Ecol Evol. doi: https://doi.org/10.1111/2041-210X.12593

The paper, which is freely available, describes the package and the data it wraps in detail. Rather than rehash the information here, we will use this post to briefly introduce the goals of the package and thank some of the people that helped it come to be.

🔗 What data does Open Tree have and how can rotl help you get it?

The Open Tree of Life combines knowledge from thousands of scientific studies to produce a single source of information about the relationships among all species on earth. In addition to storing the trees and taxonomies that go into this project, the Open Tree provides a “synthesis tree” that represents this combined knowledge. The Open Tree data can be accessed via the web page linked above, and through an API.rotl takes advantage of this API to give R users the ability to search for phylogenetic information and import the results into their R sessions. The imported data can then be used with the growing ecosystem of packages for phylogenetic and comparative biology in R.

🔗 A small example

Using rotl to get a tree for a set of names is usually a two step process. You need to start by matching the set of names you have to records in the Open Tree database, and obtain unique IDs for each one. The function tnrs_match_names handles this task:

library(rotl)
apes <- c("Pongo", "Pan", "Gorilla", "Hoolock", "Homo")
resolved_names <- tnrs_match_names(apes)
resolved_names

##   search_string unique_name approximate_match ott_id is_synonym flags
## 1         pongo       Pongo             FALSE 417949      FALSE
## 2           pan         Pan             FALSE 417957      FALSE
## 3       gorilla     Gorilla             FALSE 417969      FALSE
## 4       hoolock     Hoolock             FALSE 712902      FALSE
## 5          homo        Homo             FALSE 770309      FALSE
##   number_matches
## 1              1
## 2              1
## 3              1
## 4              1
## 5              1

Once you have a good set of unique IDs, you will want to fetch a tree. You can use ott_id to extract just the IDs from the table returned by tnrs_match_names and pass those to tol_induced_subtree to fetch a tree with theses species in the tips

tr <- tol_induced_subtree(ott_ids=ott_id(resolved_names))
plot(tr)

img

The tree is an object of the ape class phylo, which is used by most other phylogeny packages in R. There are two vignettes describing how the fetched trees can be paired with other data and used to reproduce a comparative analysis. Another vignette provides and introduction to the package and includes an FAQ.

🔗 Thanks

We’d like to thank the Open Tree project for providing such an amazing resource, and for producing a powerful and well-documented API. rotl was initially developed as part of the Open Tree’s hackathon in Ann Arbor in 2014. We’d like to thank the Open Tree project for inviting us to the meeting and all participants of that meeting for discussions and their feedback on rotl. The package was reviewed as part of the rOpenSci onboarding process, and we’d like to thank Scott Chamberlain for the time and expertise that went into this review and improved the package.