Where the tree of life is empirically resolved, and where it is not: an open atlas of species-level phylogenies and their archival uncertainty

This Zenodo deposit is the methodological and provenance archive for the Phylogenetic Atlas, a curated standardised collection of 264 empirical species-level phylogenies spanning 62 partitions of the tree of life. It contains the recipe (metadata, per-tree provenance, fetching code, standardisation pipeline, and the R package) that lets any third party reconstruct the atlas from original sources. The standardised Newick tree files themselves are not redistributed here — they live on the live atlas website (https://franciscorichter.github.io/phylo-species-atlas/) and remain canonically archived in their original publication repositories (each tree's download_url is recorded in metadata/data_provenance.csv). Headline numbers 264 standardised trees across 62 partitions (26 with shipped canonical + 36 unrepresented) 49 source datasets from 47 taxonomic groups spanning Bacteria, Archaea, and Eukaryota 637,619 standardised labels (493,978 eukaryotic + 143,641 prokaryotic GTDB clusters) 246 of 264 trees time-calibrated (93%) Archival uncertainty: only 7 of 28 non-Condamine dated source trees preserve recoverable per-node uncertainty Coverage estimates (three explicit sensitivity bounds) C0 (permissive headline): 23.7% of described eukaryotic species — all retained canonicals in full C1 (strict molecular-only on TACT-imputed trees): ~10% — molecular tips only for seed plants, squamates, sharks C2 (strictest non-imputed, non-OTL-pipeline): ~6.6% — exclude full-OTL-pipeline trees + count only molecular tips elsewhere All bounds computed from per-tree molecular_tip_fraction in metadata/S5_per_tree_provenance.csv. Three-category tree classification Direct empirical phylogenies — inferred from molecular or genomic data in a primary study. Peer-reviewed empirical syntheses — clade-bounded standalone trees with a primary methodology paper (e.g., McTavish et al. 2025 birds, Smith & Brown 2018 seed plants, Chesters 2017 insects). Excluded: the live ~2.4M-tip Open Tree of Life synthesis product. What's in the zip phylo-species-atlas-v1.0.3-zenodo.zip contains: README.md, CITATION.cff, LICENSE (CC-BY-4.0), CHANGELOG.md, manifest.json, metadata/ (data_provenance.csv with per-tree source URLs, dictionary.csv with 637,619 labels, S5/S6 provenance tables, per-partition audits), and code/ (standardisation pipeline, atlas scripts, manuscript figure-generation scripts, phyloatlas R package source v0.1.0). How to reconstruct standardised trees Read metadata/data_provenance.csv — every tree's original source URL is in the download_url column. Use code/pipeline/ to download from those URLs, clean tip labels, and rebuild standardised Newick files + dictionary. Or load any tree directly from R using the phyloatlas package. Or download pre-standardised trees from the live atlas website. Manuscript This deposit accompanies: Richter, F. (2026). Where the tree of life is empirically resolved, and where it is not: an open atlas of species-level phylogenies and their archival uncertainty. Concurrent bioRxiv preprint. Cross-references Live atlas: https://franciscorichter.github.io/phylo-species-atlas/ GitHub: https://github.com/franciscorichter/phylo-species-atlas Concept DOI (always resolves to latest): https://doi.org/10.5281/zenodo.20127157