rsomics-tree-wrfd

Weighted Robinson-Foulds distance between two phylogenetic trees: the branch-length-vector distance over matched bipartitions. Each bipartition (or, on rooted trees, each clade) shared between the two trees contributes the difference of its branch lengths; a bipartition present in only one tree is given length 0 in the other. The default cityblock reduction is the original weighted RF distance; the euclidean reduction is the Kuhner-Felsenstein branch score.

rsomics-tree-wrfd t1.nwk t2.nwk [--metric cityblock|euclidean] [--no-tips]

The companion topology overlaps compare_biparts and compare_subsets — the normalized symmetric difference of the bipartition or subset sets, ignoring branch lengths — are available as --compare biparts and --compare subsets. They are the same "compare two trees" operation at a different weighting, which is why they live here rather than in a separate crate.

Kept separate from rsomics-tree-rfd: weighted and unweighted RF are distinct user-facing operations with different output meaning (branch-length divergence vs a topology count).

Examples

# Weighted RF (cityblock)
rsomics-tree-wrfd a.nwk b.nwk

# Kuhner-Felsenstein branch score, terminal branches excluded
rsomics-tree-wrfd a.nwk b.nwk --metric euclidean --no-tips

# Normalized bipartition overlap (compare_biparts)
rsomics-tree-wrfd a.nwk b.nwk --compare biparts

Rooting is inferred from the first tree (rooted iff its root has exactly two children), matching scikit-bio. Override with --rooted / --unrooted. Only taxa shared between the two trees are considered.

Origin

This crate is an independent Rust reimplementation of scikit-bio's TreeNode.compare_wrfd, compare_biparts and compare_subsets, informed by reading the scikit-bio source (BSD-3-Clause), which we are permitted to read and cite:

• Robinson, D. F., & Foulds, L. R. (1979). Comparison of weighted labelled trees. Combinatorial Mathematics VI, 119-126.
• Kuhner, M. K., & Felsenstein, J. (1994). A simulation comparison of phylogeny algorithms under equal and unequal evolutionary rates. Molecular Biology and Evolution, 11(3), 459-468. DOI: 10.1093/oxfordjournals.molbev.a040126

Values are reproduced to within ~1e-9 of scikit-bio (the weighted reduction is a deterministic f64 sum; bipartition matching is exact set logic).

License: MIT OR Apache-2.0. Upstream credit: scikit-bio https://scikit-bio.org (BSD-3-Clause).