Skip to content
DraconiticResearch
Search Sources API
197,788 studies31 sourcesUpdated 22 min ago
PreprintCC BYbioRxiv

An unconstrained Euclidean parametrisation of phylogenetic tree space

  1. 1University of Warwick

Abstract

Phylogenetic trees combine a discrete topology with continuous branch lengths, which makes them difficult to work with. We introduce a recursive contour-grafting construction that encodes both components using unconstrained real-valued coordinates. After fixing an order of the leaves, each new leaf is attached at a point selected by a contour traversal of the tree constructed so far. Moving a graft coordinate continuously along the contour enables changes of topology when internal branches contract to zero. For trees with n labelled leaves, we apply the construction to rooted ultrametric trees, rooted trees with unrestricted branch lengths, and unrooted trees, obtaining parametrisations of dimensions n-1, 2n-2, and 2n-3, respectively. We show that the resulting construction functions are surjective and continuous onto the corresponding spaces of phylogenetic trees. We also determine the number of representations of a generic tree and calculate the Jacobian of each parametrisation. These results provide explicit change-of-variables formulae between Euclidean coordinates and the phylogenetic tree spaces, and we show how probability distributions on tree space can be sampled using generic algorithms that target unconstrained Euclidean coordinates. This approach paves the way for the development and application of new methodology in Bayesian phylogenetic inference.

Related

  1. Basis statistics for phylogenetic tree shape: tree balance formulas beyond expectations
    Daniel J. Bauman, Noah A. Rosenberg
    arXivPreprintCC BY 4.0Full textEvolutionary biologyProbabilityarXiv
  2. The nearest neighbor interchange proximity measure for the tree-basedness of rooted binary phylogenetic networks
    Martin Frohn, Takatora Suzuki
    arXivPreprintarXiv licenceFull textEvolutionary biologyOptimisation & controlarXiv
  3. The parental parsimony problem on binary, tree-child phylogenetic networks
    Martin Frohn
    arXivPreprintarXiv licenceFull textEvolutionary biologyOptimisation & controlarXiv
  4. Finding Tree-Like Substructures in Phylogenetic Networks: ILP Approaches and Their Application
    Takatora Suzuki
    arXivPreprintCC BY 4.0Full textEvolutionary biologyarXiv
  5. Graph construction in QUBO-based recursive phylogenetic tree reconstruction
    Yoshiki Kanazawa, Takahiko Koyama
    arXivPreprintarXiv licenceFull textEvolutionary biologyQuantum physics & quantum informationarXiv
BibTeX
@misc{didelot2026unconstrained,
  title = {{An unconstrained Euclidean parametrisation of phylogenetic tree space}},
  author = {X. Didelot},
  howpublished = {bioRxiv},
  year = {2026},
  month = oct,
  doi = {10.64898/2026.10.14.759122},
  url = {https://www.biorxiv.org/content/10.64898/2026.10.14.759122v1},
  note = {CC BY},
}
RIS
TY  - UNPB
TI  - An unconstrained Euclidean parametrisation of phylogenetic tree space
AU  - X. Didelot
PY  - 2026
DA  - 2026/10/14
PB  - bioRxiv
DO  - 10.64898/2026.10.14.759122
UR  - https://www.biorxiv.org/content/10.64898/2026.10.14.759122v1
AB  - Phylogenetic trees combine a discrete topology with continuous branch lengths, which makes them difficult to work with. We introduce a recursive contour-grafting construction that encodes both components using unconstrained real-valued coordinates. After fixing an order of the leaves, each new leaf is attached at a point selected by a contour traversal of the tree constructed so far. Moving a graft coordinate continuously along the contour enables changes of topology when internal branches contract to zero. For trees with n labelled leaves, we apply the construction to rooted ultrametric trees, rooted trees with unrestricted branch lengths, and unrooted trees, obtaining parametrisations of dimensions n-1, 2n-2, and 2n-3, respectively. We show that the resulting construction functions are surjective and continuous onto the corresponding spaces of phylogenetic trees. We also determine the number of representations of a generic tree and calculate the Jacobian of each parametrisation. These results provide explicit change-of-variables formulae between Euclidean coordinates and the phylogenetic tree spaces, and we show how probability distributions on tree space can be sampled using generic algorithms that target unconstrained Euclidean coordinates. This approach paves the way for the development and application of new methodology in Bayesian phylogenetic inference.
ER  -