v1.0.8 — turtle chronogram canonical; MEE integrity pass

Data-integrity release accompanying the MEE resubmission.

• Turtles: the standardized canonical is now the 274-tip ultrametric Thomson 2021 MCC chronogram (crown ~212 Ma), replacing a 287-tip substitution tree that failed time-calibration integrity. Reconciled across metadata.csv, data_provenance, data_estimates, and Table S5.
• Fungi: manuscript reconciled to the 1,602 species-level shipped count (one tip per species; the 1,672 raw-tip tree carries 22 duplicate-species tips).
• Carries the cacti/conifers/solanaceae dating curation and the corrected Condamine DOI (10.1111/ele.13382).
• 264 trees, 247 time-calibrated, 62 partitions; every dated tree ships a verified ultrametric chronogram.
• TimeTree-of-Life remains cited, not redistributed.

Concept DOI: 10.5281/zenodo.20127157 (always resolves to the latest version).

v1.0.7 — restore species-name dictionary

v1.0.7 — Restore species-name dictionary

The v1.0.6 TimeTree de-redistribution over-pruned the flat species-name
dictionary. A list of Linnaean species names is not TimeTree's tree, so this was
out of scope. This release restores the dictionary to its full 637,619
standardized labels.

The TimeTree-of-Life (Hedges et al. 2015) tree, its name shard, and its
tree-membership rows remain removed — it stays cited, not redistributed
(available from timetree.org).

v1.0.6 — data-integrity corrections + TimeTree de-redistribution

v1.0.6 — Data-integrity corrections + TimeTree de-redistribution

Time-calibration fixes

• Shipped the published time-calibrated chronograms for 8 partitions that previously shipped substitution/topology trees: turtles, birds, cetaceans, cacti, orchids, parrots, solanaceae, conifers.
• Crustaceans now ships Wolfe et al. (2019)'s CIR chronogram; diatoms reconciled to Nakov et al. (2018).
• Bryophytes reclassified as undated (no time-calibrated tree was deposited).
• Time-calibrated trees: 247 of 264.
• Turtles canonical updated to Thomson et al. (2021)'s dated MCC (274 species).

Citation / metadata corrections

• Provenance and DOIs corrected for acari, tunicates, other_hexapods, bryozoa, nematodes.

Licensing

• The TimeTree-of-Life (Hedges et al. 2015) is now cited as a cross-cutting reference but not redistributed, at the TimeTree project's request — available from https://timetree.org.

The 264-tree decomposition (218 Condamine + 25 partition-canonical + 21 sub-clade/reference) and the 62 partitions are unchanged.

v1.0.5 — consistency corrections

Consistency corrections aligning the deposit with the submitted manuscript (MEE-26-05-507; bioRxiv v2).

• Coverage: strictest sensitivity bound C2 corrected to ~5.4% (was 6.6%); Catalogue of Life match rate 83.2%; recoverable archival uncertainty 7 of 24 non-Condamine dated source trees.
• Fish: coverage reported as the shipped 11,638-tip molecular subset (32.2%), distinct from the 31,516-tip imputed source tree (87.5%).
• New deposited file: audits/table_s6_alternative_trees.csv (19 sub-clade + 2 cross-cutting reference trees), auto-generated by site/build.py.
• Partitions: added the 62nd partition (Other vertebrates) to data_estimates.csv.
• Tree decomposition corrected: 218 Condamine family-level (incl. the crocodilians canonical, itself a Condamine tree) + 25 non-Condamine canonical + 21 sub-clade/reference. The previous "26 + 218 + 20" double-counted the crocodilians canonical.

v1.0.4 — data corrections + canonical succession (Supplementary Table S7)

v1.0.4 (2026-05-31)

Concept DOI: https://doi.org/10.5281/zenodo.20127157 (always resolves to latest)
Version DOI: auto-minted by Zenodo upon publication of this release.

Data corrections — 4 partitions reclassified dated → undated

The atlas's per-partition override (site/data/partitions/<slug>/info.yaml)
or curation heuristic had recorded these four canonical trees as time-calibrated.
Verification against the source papers confirms they ship phylogenomic ML
(or Bayesian) topologies without time calibration:

Partition	Source paper	What's actually deposited
Bivalves	Pfeiffer et al. 2019 (MPE)	Anchored hybrid enrichment ML topology, no chronogram
Bryozoa	Orr et al. 2022 (Sci Adv)	SI_Fig2 ML topology pruned to species; chronogram described in Fig 2 but never deposited
Crustaceans	Wolfe et al. 2019 (PRSB)	410-locus Bayesian topology, no chronogram
Gastropods	Zapata et al. 2014 (PRSB)	Transcriptome ML backbone, no chronogram

data_provenance.csv (the source of truth) was already correct for all four;
the drift was in site/data/partitions/bryozoa/info.yaml (and the per-tree
provenance ledger in the paper). All four now show consistent dated: false
in both source CSVs and the derived site/data.json.

New artefacts under audits/

• table_s5_canonical_tree_provenance.csv — 27 rows; per-tree provenance
  ledger (the manuscript Supplementary Table S5). Four rows carry
  [CORRECTED 2026-05-31] notes documenting the dated→undated change.
• table_s7_canonical_succession.csv — NEW, 95 rows. Historical
  succession of canonical species-level trees per partition from 2006 to
  2026. Each row records publication year, citation key, Crossref-resolvable
  DOI, tip count, molecular-tip fraction, dating code, OTL pipeline
  involvement, supersession chain, and a canonical_in_atlas flag (yes
  for the 26 rows whose DOI matches the atlas-shipped canonical; no for
  the 69 predecessors and non-adopted post-canonical alternatives).
• canonical_succession_audit.md — per-partition coverage report.
  Confidence breakdown of accepted rows: 41 high / 53 medium / 1 low.
  1 entry rejected.

New manuscript figure (now in code/figures/)

• fig2_archival_uncertainty.py + code/figures/outputs/figure_archival_uncertainty.{png,pdf} — Figure 2 of the manuscript: by-year stacked-bar
  of canonical source trees, classed by HPD-recoverability. Headline:
  7 of 26 atlas canonicals (27%) preserve a recoverable divergence-time
  distribution; 10 are point-dated only; 9 are undated phylogenomic or
  supertree topologies. Reads from audits/table_s5_canonical_tree_provenance.csv.

Numeric claims updated downstream in the manuscript

• Time-calibrated trees: 246 / 264 (93%) → 242 / 264 (92%)
• Crown-age uncertainty audit denominator: 28 → 24 non-Condamine dated source trees
• 7 / 28 (25%) recoverable HPDs → 7 / 24 (29%)

Other

• code/figures/fig1_dotplot.py and fig2_crown_age.R updated to render the
  blue / green / amber diamond convention introduced in the manuscript.

v1.0.3 — title alignment, LICENSE, three-category framework, sensitivity bounds

v1.0.3 (2026-05-30)

Zenodo release: https://doi.org/10.5281/zenodo.20467013 (version DOI)
Concept DOI: https://doi.org/10.5281/zenodo.20127157 (always resolves to latest)

Architecture and metadata alignment with the accompanying manuscript.

Manuscript title

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

Architecture shift — what the Zenodo deposit now archives

The v1.0.3 Zenodo deposit holds the methodological recipe (metadata, per-tree provenance, fetching/standardisation code, R package source). The standardised Newick tree files themselves are NOT redistributed via Zenodo — they live on this GitHub repo + the live atlas website + their original publication repositories. See staging/zenodo/CHANGELOG.md in the v1.0.3 deposit for rationale.

Highlights

• Three-category tree classification (direct empirical / peer-reviewed empirical synthesis / excluded live OTL synthesis product) operationalised in S5 per-tree provenance.
• Three explicit sensitivity bounds on eukaryotic coverage:
  • C0 = 23.7% (permissive headline, all retained canonicals in full)
  • C1 ≈ 10% (strict molecular-only on the three TACT-imputed trees)
  • C2 ≈ 6.6% (strictest: exclude full-OTL-pipeline trees + count only molecular tips elsewhere)
• Archival-uncertainty audit retained: only 7 of 28 non-Condamine dated source trees preserve recoverable per-node uncertainty.

Repo-side changes (this release)

• CITATION.cff: new manuscript title, version 1.0.3, concept DOI 10.5281/zenodo.20127157
• README.md: rewritten header with At-a-glance table + three-category framework
• LICENSE: CC BY 4.0 full text
• code/figures/: complete manuscript figure-generation scripts
• code/pipeline/: standardisation pipeline + supporting scripts

Corrections

• Mammals molecular_tip_fraction: was 0.88 (unsupported); now 0.69 (strict type-1 molecular per Upham 2019).

Cross-references

• Zenodo v1.0.3: https://doi.org/10.5281/zenodo.20467013
• Concept DOI: https://doi.org/10.5281/zenodo.20127157
• Live atlas: https://franciscorichter.github.io/phylo-species-atlas/
• bioRxiv preprint: forthcoming

v1.0.2 — Initial Release

v1.0.2 — Initial Release

What's Included

• standardized/trees/: 264 Newick files with numeric tip labels
• standardized/dictionary.csv: Global species dictionary (641,763 entries)
• standardized/metadata.csv: Per-tree metadata (group, source, tips, dated status)
• standardized/full_mapping.csv: Original label → standardized name → ID → group
• code/: Curation and standardization scripts

Coverage Summary

Metric	Value
Datasets	49
Taxonomic groups	47
Unique standardized labels	641,763
Eukaryotic species	498,149
Prokaryotic genome clusters	143,614
Dated phylogenies	30
Undated phylogenies	19

v1.0.1 — Initial Release

v1.0.1 — Initial Release

What's Included

• standardized/trees/: 264 Newick files with numeric tip labels
• standardized/dictionary.csv: Global species dictionary (641,763 entries)
• standardized/metadata.csv: Per-tree metadata (group, source, tips, dated status)
• standardized/full_mapping.csv: Original label → standardized name → ID → group
• code/: Curation and standardization scripts

Coverage Summary

Metric	Value
Datasets	49
Taxonomic groups	47
Unique standardized labels	641,763
Eukaryotic species	498,149
Prokaryotic genome clusters	143,614
Dated phylogenies	30
Undated phylogenies	19