For evolutionary biologists, ecologists, and computational researchers, building a phylogeny at large taxonomic scales has traditionally meant assembling information scattered across hundreds or thousands of publications. They could spend weeks assembling one by hand from scattered papers, or settle for a resource that covered only a slice of life with inconsistent taxon names and no shared identifiers across studies. Phylogenetic knowledge was locked in individual publications, each describing a different branch of the tree in a different file format, making it impossible to combine and keep current.
Open Tree of Life (OTL) was built to address this problem by providing an openly accessible, machine-readable synthesis of evolutionary relationships across the tree of life. It is best understood as a topological scaffold, not a finished, time-calibrated phylogeny. It has become exceptionally useful for retrieving relationships, linking taxonomic information, identifying phylogenetic conflict, and enabling analyses at scales that would otherwise be difficult to approach. But researchers must take additional steps when their analysis depends on branch lengths or divergence times.
What Open Tree of Life Is and Why It Was Built
OTL was launched in 2012 under National Science Foundation (NSF) funding as part of the Assembling the Tree of Life initiative, built specifically to solve that assembly problem. The first full synthetic tree appeared in 2015, when Hinchliff and colleagues published a synthesis in PNAS combining 484 source phylogenies drawn from 3,062 studies — resolving relationships for roughly 38,000 tips directly from phylogenetic data and placing the remaining 2.3 million or so tips using the Open Tree Taxonomy.
The engine behind this is a synthesis algorithm called propinquity. It loads all sources of phylogeny, along with the taxonomy, into a single graph of life, then traverses that graph to build a synthetic tree that favours well-supported, expert-curated phylogenies over taxonomic guesswork. When source trees disagree, OTL doesn’t quietly pick a winner — it flags the conflict, so a researcher can see exactly which studies support or contradict any given node.
By the current synthesis, v15.1 (released July 2024), the tree had grown to 2,384,572 tips, of which 129,778 are informed directly by phylogenetic evidence. The rest are placed by taxonomy alone. That gap between total tips and phylogenetically supported tips is arguably the single most important number to understand before using OTL for anything analytical.
What Does the Open Tree of Life Enable?
1. Rapid access to large-scale phylogenetic topologies
The most immediate advantage of OTL is its scale and accessibility. The resource can be accessed through tools including the rotl R package, the opentree Python package, REST APIs, and the OTL web interface. For groups with substantial representation in the underlying source phylogenies, the resulting topology can provide a well-resolved representation of evolutionary relationships.
For researchers working across thousands of species, this is more than a convenience. It changes the practical scale at which comparative and macroevolutionary questions can be formulated.
2. Phylogenetic diversity and conservation analyses
The scale of OTL makes large-scale analyses of evolutionary diversity possible. Measures such as phylogenetic diversity (PD) and evolutionary distinctiveness (ED) require information about relationships among large numbers of species.
Conservation prioritisation frameworks can similarly benefit from broad phylogenetic coverage. Global analyses involving millions of species would be extremely difficult to conduct using individually assembled literature trees. OTL provides the underlying topological framework that makes these questions computationally approachable — although analyses requiring evolutionary distances or time must account for its lack of branch lengths.
3. Detection of phylogenetic conflict
One of OTL’s less obvious strengths is that synthesis does not merely produce a single tree. Its underlying infrastructure retains information about the source studies contributing to relationships. Researchers investigating a group with disputed evolutionary relationships can therefore examine which published phylogenies support a node and which conflict with it. The connection to the Phylesystem data store provides a route back to the source studies.
For a researcher, this can turn what would otherwise be a manual literature-review exercise into a queryable investigation of phylogenetic evidence.
4. Cross-database taxonomic linking through OTT
The Open Tree Taxonomy (OTT) adds another layer of value. OTT integrates taxonomic information from sources including NCBI Taxonomy, GBIF, IRMNG, and Index Fungorum. OTT identifiers can therefore serve as practical anchors when connecting information across biodiversity databases.
A researcher can use an OTT identifier to connect a taxon with occurrence information, sequence repositories, or other biological datasets where corresponding taxonomic identifiers are available. Both rotl and opentree provide functionality for taxonomic name matching and cross-database lookup.
This makes OTL relevant not only as a phylogenetic resource but also as part of a broader biodiversity informatics workflow.
The Limitations of OTL
OTL gives you topology, not time. That’s not an oversight but a deliberate response to a real constraint. When you combine phylogenies built from different data types, different taxon sets, and different calibration schemes, merging their branch lengths into something meaningful simply isn’t possible.
This matters because so much downstream analysis depends on branch lengths in units of evolutionary time. Trait evolution models, diversification rate methods like BAMM or RPANDA, and distance-based community phylogenetics all need them. OTL’s synthetic tree can’t be plugged directly into any of these. The standard workaround is a three-step process: pull the OTL topology for your group, apply divergence time estimation using a tool like DateLife or treePL, then run your analysis on the dated result. A 2026 preprint on assembling a fully dated tree of life took a run at automating this at scale, developing interpolation algorithms to get around the fact that standard dating approaches become computationally impossible once you’re working with millions of nodes.
The taxonomy-placed majority is the other constraint worth sitting with. Of the 2,384,572 tips in v15.1, only 129,778 owe their position to actual phylogenetic evidence. Everything else is placed by taxonomic assignment, meaning its reliability tracks the underlying taxonomy rather than a published molecular tree. For groups like mammals or birds, where taxonomy and phylogeny mostly agree, this rarely causes problems. For invertebrates, fungi, and protists, where the two regularly diverge, it’s a meaningful source of uncertainty.
None of this is really OTL’s fault. The synthesis can only incorporate what researchers actually deposit, and a large share of published phylogenetic work still ends up in supplementary PDFs or formats that can’t be machine-parsed, with tip labels never mapped to OTT identifiers. The Phylesystem store currently holds more than 1,216 studies, which is a real number, but a fraction of everything that’s been published. That’s a community data-sharing problem more than a design flaw, but it directly caps how resolved the synthetic tree can become.
How Researchers Correctly Use OTL
For most downstream analyses, a sensible workflow has four stages.
First, retrieve the topology. Use rotl::tol_induced_subtree() or the corresponding functionality in the opentree Python package to obtain the OTL topology for the taxa in your analysis.
Second, inspect the evidence. Check the support and conflict information associated with important nodes. This is particularly important if your conclusions depend on whether a trait evolved once or multiple times, or on relationships among closely related taxa.
Third, add divergence times when required. If your analysis requires evolutionary distances or time, apply an appropriate dating method or use a suitable dated resource rather than treating the OTL topology itself as time-scaled.
Fourth, use OTT identifiers for taxonomic integration. Name matching through tnrs_match_names() can help map species lists to OTT identifiers, which can then be connected with other biodiversity databases. Taxonomically complex groups should still be checked manually, because automated name matching is not infallible.
Reproducibility also requires recording the synthesis version, API version, and access date. Cite the original synthesis paper as well as the specific OTL version used in your analysis.
How Is OTL Changing?
OTL is not a static tree. Each new synthesis can incorporate additional source phylogenies deposited through Phylesystem. The increase from approximately 38,000 phylogenetically informed tips in the first 2015 synthesis to 129,778 in version 15.1 demonstrates substantial progress, although it remains a small proportion of the total number of tips.
One particularly important area of development is scalable divergence-time estimation. A 2026 preprint described interpolation approaches designed to address the computational challenge of dating a tree at the scale of millions of species. If such approaches mature and become integrated with OTL, the distinction between a massive topological scaffold and a fully dated tree could become less restrictive for large-scale analyses.
Integration with biodiversity databases is another natural direction. As shared taxonomic identifiers become more deeply embedded across resources such as GBIF and NCBI, OTT IDs could become increasingly useful as anchors connecting phylogenetic, occurrence, sequence, and trait information.
Automated curation may also improve the rate at which published evidence enters the synthesis. NLP-based systems capable of extracting phylogenetic trees from publications and mapping taxon names to OTT identifiers could reduce the manual burden currently involved in curating source data.
The Bottom Line
Open Tree of Life remains the only resource that gives researchers free, immediate, machine-readable access to a tree spanning all 2.3 million named species. For pulling topologies, surfacing phylogenetic conflict, linking taxonomy across databases, and enabling macroevolutionary questions at global scale, it’s infrastructure the field has come to depend on.
OTL doesn’t hand researchers a finished phylogeny. It hands them the best current summary of what the community collectively knows about evolutionary relationships, in a queryable, openly licensed form that improves with every study that gets deposited. It’s worth more than it sounds.
Call to Action
For researchers working across large bodies of scientific literature and biodiversity data, the challenge is increasingly not finding individual papers or datasets, but connecting the evidence they contain. Axy is built for that gap: a scientific knowledge graph that organises relationships between literature, taxonomic entities, datasets, and research concepts — where every link is auditable, conflicting evidence is preserved rather than flattened, and every change passes human review.
If your work involves phylogenetics, biodiversity informatics, or literature-scale discovery, apply to join below and help build the shared, connected map of science rather than just use another tool.
