# Proposed GO taxon constraints, collected by the Taxon Pathway Variance project. # One entry per (term, constraint, taxon). Every GO and NCBITaxon id is verified # (OLS / QuickGO / NCBITaxon), never written from memory. # # Fields: # go_term: {id, label} # constraint: never_in_taxon | only_in_taxon # taxon: {id, label} # existing_go_constraint: what GO already asserts on the term or its ancestors: null (none), # or a list of {constraint, taxon} (TPV-004 onward) # conservative_alternative: optional narrower or weaker option, with its rationale # evidence: list of {reference_id, supporting_text} (verbatim quotes) # blocks: annotations the constraint would remove (gene, protein, organism, # evidence, and source_node for IBA or source_id for other sources) # would_wrongly_block: members of the taxon that do have the presupposed component, or [] # prerequisites: ontology changes needed before the constraint can be applied: # {kind, term, change, reason, status: NOT_REQUESTED | REQUESTED | DONE, # notes (optional, e.g. when the prerequisite has been overtaken)} # in_repo_agreement: reviews in this repo that reach the same conclusion # notes: free text (counts, caveats, matching upstream fixes) # status: PROPOSED | SUBMITTED | ACCEPTED | REJECTED # submission: once SUBMITTED: {tracker_issue (URL), submitted (date), # resolved (date), resolution (as GO implemented it), commits # (URLs), pull_requests (URLs), notes} # source: where in this repo the case was found proposals: - id: TPV-001 go_term: id: GO:0007259 label: cell surface receptor signaling pathway via JAK-STAT constraint: never_in_taxon taxon: id: NCBITaxon:6231 label: Nematoda conservative_alternative: taxon: id: NCBITaxon:6237 label: Caenorhabditis rationale: The direct statement of JAK absence is for C. elegans. PANTHER's JAK family (PTHR45807) has no nematode leaf and the profile matches only five nematode fragments, and has no member in C. briggsae or Pristionchus pacificus, so the phylum-wide claim is an inference from genome coverage, not a statement for every nematode. existing_go_constraint: null evidence: - reference_id: PMID:28874466 supporting_text: There is no conserved homolog of the JAK kinases in C. elegans blocks: - gene: sta-1 protein: UniProtKB:Q9NAD6 organism: Caenorhabditis elegans evidence: IBA source_node: PANTHER:PTN000927860 - gene: sta-2 protein: UniProtKB:Q20977 organism: Caenorhabditis elegans evidence: IBA source_node: PANTHER:PTN000927860 - gene: T12G3.2 protein: UniProtKB:D3YT30 organism: Caenorhabditis elegans evidence: IBA source_node: PANTHER:PTN000927860 - gene: sta-1 protein: UniProtKB:Q61AP6 organism: Caenorhabditis briggsae evidence: IBA source_node: PANTHER:PTN000927860 - gene: sta-2 protein: UniProtKB:B0K0A6 organism: Caenorhabditis briggsae evidence: IBA source_node: PANTHER:PTN000927860 - protein: UniProtKB:A8X0M7 organism: Caenorhabditis briggsae evidence: IBA source_node: PANTHER:PTN000927860 - protein: UniProtKB:A0A2A6BZA0 organism: Pristionchus pacificus evidence: IBA source_node: PANTHER:PTN000927860 - protein: UniProtKB:A0A2A6CKG1 organism: Pristionchus pacificus evidence: IBA source_node: PANTHER:PTN000927860 - protein: UniProtKB:A0A914YCL6 organism: Panagrolaimus sp. evidence: IEA (TreeGrafter) source_node: PANTHER:PTN001562978 note: 203-aa Ser/Thr kinase fragment misplaced into the JAK family would_wrongly_block: [] in_repo_agreement: - genes/worm/sta-1/sta-1-ai-review.yaml (REMOVE) - genes/worm/sta-2/sta-2-ai-review.yaml (REMOVE) - interpro/panther/PTHR11801/PTHR11801-review.yaml (GO:0007259 SUBFAMILY_ONLY, nematode subfamilies excluded; PTN000927860 TOO_DEEP) - interpro/panther/PTHR45807/PTHR45807-review.yaml (member exception for A0A914YCL6) status: ACCEPTED submission: tracker_issue: https://github.com/geneontology/go-ontology/issues/32689 submitted: 2026-10-05 resolved: 2026-11-03 resolution: >- Added as is_a: neverin:6231 (Nematoda) on GO:0007259 in src/ontology/go-edit.obo, with a term_tracker_item pointing to #32689. The OWL diff on go-ontology#32857 confirms the added axiom on GO_0007259 ("SubClassOf never in Nematoda"). commits: - https://github.com/geneontology/go-ontology/commit/98389d1bd9ea9c7e60b5a4420b7cf61c9ce5b0be pull_requests: - https://github.com/geneontology/go-ontology/pull/32857 source: modules/c_elegans_jak_independent_stat_signaling.yaml - id: TPV-002 go_term: id: GO:0046425 label: regulation of cell surface receptor signaling pathway via JAK-STAT constraint: never_in_taxon taxon: id: NCBITaxon:6231 label: Nematoda conservative_alternative: taxon: id: NCBITaxon:6237 label: Caenorhabditis rationale: As for TPV-001. existing_go_constraint: null notes: Stated separately from TPV-001 because it is not established that GO's taxon-constraint reasoning propagates a never_in_taxon on GO:0007259 to terms that regulate it. Covers the positive and negative regulation children (GO:0046427, GO:0046426) by is_a. evidence: - reference_id: PMID:28874466 supporting_text: There is no conserved homolog of the JAK kinases in C. elegans blocks: - gene: gei-17 protein: UniProtKB:Q94361 organism: Caenorhabditis elegans term: GO:0046426 evidence: IBA source_node: PANTHER:PTN000845825 - protein: UniProtKB:A8XFG1 organism: Caenorhabditis briggsae term: GO:0046426 evidence: IBA source_node: PANTHER:PTN000845825 - protein: UniProtKB:A0A8R1U6M4 organism: Pristionchus pacificus term: GO:0046426 evidence: IBA source_node: PANTHER:PTN000845825 - protein: UniProtKB:A0A8R1Z6J1 organism: Pristionchus pacificus term: GO:0046426 evidence: IBA source_node: PANTHER:PTN001057235 - description: about 100 nematode TreeGrafter IEAs for GO:0046426 evidence: IEA would_wrongly_block: [] in_repo_agreement: - genes/worm/gei-17/gei-17-ai-review.yaml (REMOVE) - interpro/panther/PTHR10782/PTHR10782-review.yaml (member exception for GEI-17, PTN000845825 NEEDS_PRUNING) status: ACCEPTED submission: tracker_issue: https://github.com/geneontology/go-ontology/issues/32689 submitted: 2026-10-05 resolved: 2026-11-03 resolution: >- Added as is_a: neverin:6231 (Nematoda) on GO:0046425, with a term_tracker_item pointing to #32689. The same commit renamed GO:0046425 to "regulation of cell surface receptor signaling pathway via JAK-STAT" (old label kept as an exact synonym) and renamed GO:0046426 and GO:0046427 likewise, under go-ontology#14167. The regulation children carry no constraint of their own and inherit this one by is_a. go_term.label above is the current label; the proposal was filed under the old one. commits: - https://github.com/geneontology/go-ontology/commit/98389d1bd9ea9c7e60b5a4420b7cf61c9ce5b0be pull_requests: - https://github.com/geneontology/go-ontology/pull/32857 source: interpro/panther/PTHR10782/PTHR10782-review.yaml - id: TPV-003 go_term: id: GO:0007259 label: cell surface receptor signaling pathway via JAK-STAT constraint: only_in_taxon taxon: id: NCBITaxon:33208 label: Metazoa conservative_alternative: constraint: never_in_taxon taxon: id: NCBITaxon:33083 label: Dictyostelia rationale: Blocks only the slime-mould STAT IBAs. Metazoa is preferred because JAKs arise in animals (PANTHER roots PTHR45807 at Eumetazoa and places a JAK in Trichoplax), so any non-animal JAK-STAT annotation is suspect. Bilateria would be tighter but would exclude the placozoan JAK. existing_go_constraint: null evidence: - reference_id: PMID:24058787 supporting_text: The first bona fide JAK-like protein is represented in the basal metazoan porifera lineage - reference_id: PMID:16596165 supporting_text: The human kinases are in the Jak, RSK, Trio, and GCN2 families, of which only GCN2 (IfkA) is found in Dictyostelium - reference_id: PMID:22699506 supporting_text: There are no orthologs of the animal tyrosine kinases blocks: - gene: dstA protein: UniProtKB:O00910 organism: Dictyostelium discoideum evidence: IBA source_node: PANTHER:PTN000927860 - gene: dstB protein: UniProtKB:Q70GP4 organism: Dictyostelium discoideum evidence: IBA source_node: PANTHER:PTN000927860 - gene: dstC protein: UniProtKB:Q54BD4 organism: Dictyostelium discoideum evidence: IBA source_node: PANTHER:PTN000927860 - gene: dstD protein: UniProtKB:Q86I20 organism: Dictyostelium discoideum evidence: IBA source_node: PANTHER:PTN000927860 - description: further Dictyostelium STATs proteins: - UniProtKB:Q54B81 - UniProtKB:Q54GL9 - UniProtKB:Q54WB7 - UniProtKB:F0ZVN4 - UniProtKB:F1A659 - UniProtKB:F0Z7T3 organisms: - Dictyostelium discoideum (Q54B81, Q54GL9, Q54WB7) - Dictyostelium purpureum (F0ZVN4, F1A659, F0Z7T3) evidence: IBA source_node: PANTHER:PTN000927860 - description: 13 fungal, 2 plant and 1 bacterial TreeGrafter IEAs from PTHR45807 nodes evidence: IEA would_wrongly_block: - term: GO:0007260 tyrosine phosphorylation of STAT protein reason: GO:0007260 is part_of GO:0007259, but Dictyostelium STATc is tyrosine phosphorylated without a JAK, by the TKL kinase Pyk2 (PMID:22699506). GO's taxon-constraint reasoning propagates over part_of, so the accepted constraint now forbids this correct annotation (see the prerequisite and submission notes below). - term: GO:0042531 (positive regulation of tyrosine phosphorylation of STAT protein) reason: Cholera toxin subunits (UniProtKB:P01555, P01556; Vibrio cholerae) carry IDA annotations that would conflict if the constraint propagates through regulates and part_of. prerequisites: - kind: ONTOLOGY_CHANGE term: id: GO:0007260 label: tyrosine phosphorylation of STAT protein change: Remove the part_of GO:0007259 link, keeping GO:0007260 under GO:0097696 cell surface receptor signaling pathway via STAT. reason: Dictyostelium STATc is tyrosine phosphorylated without a JAK (PMID:22699506), so an only_in_taxon Metazoa constraint on GO:0007259 would wrongly forbid GO:0007260 in Dictyostelium, because it propagates over part_of. status: NOT_REQUESTED notes: >- Overtaken: GO applied the constraint without this change. GO:0007260 is already listed for obsoletion in go-ontology#27064, which would remove the conflict. If it is not obsoleted, request the re-parenting on #27064. in_repo_agreement: - interpro/panther/PTHR11801/PTHR11801-review.yaml (PTN000927860 TOO_DEEP; Dictyostelium subfamilies excluded) - interpro/panther/PTHR45807/PTHR45807-review.yaml (family rooted at Eumetazoa) - genes/DICDI/statA and genes/DICDI/statC (JAK-STAT IBA modified to GO:0097696) status: ACCEPTED submission: tracker_issue: https://github.com/geneontology/go-ontology/issues/32689 submitted: 2026-10-05 resolved: 2026-11-03 resolution: >- Added as is_a: onlyin:33208 (Metazoa) on GO:0007259, alongside the TPV-001 neverin:6231, merged in go-ontology#32861. The GO:0007260 prerequisite below was not part of the change. commits: - https://github.com/geneontology/go-ontology/commit/fed722deb9e08f121365216c338db4cc08fac3e0 pull_requests: - https://github.com/geneontology/go-ontology/pull/32861 notes: >- The constraint was first missed (only neverin:6231 landed in 98389d1). A reviewer caught this and it was added in a second commit. GO accepted the Metazoa constraint without the GO:0007260 prerequisite. The automated review on #32861 says GO:0007260 "is `part_of` GO:0007259, so it now inherits only-in Metazoa too" and that "GO:0007260 is already an obsoletion candidate, though, so this may resolve itself". GO:0007260 is listed for obsoletion in go-ontology#27064 (open), but OLS still shows it as live. If it stays live, any Dictyostelium annotation to it now violates the constraint. source: projects/IBA_REVIEW.md section 14 (statA/statC) and interpro/panther/PTHR11801/PTHR11801-review.yaml - id: TPV-004 go_term: id: GO:0030907 label: MBF transcription complex constraint: only_in_taxon taxon: id: NCBITaxon:4751 label: Fungi existing_go_constraint: - constraint: only_in_taxon taxon: id: NCBITaxon:2759 label: Eukaryota evidence: - reference_id: PMID:27162172 supporting_text: We were unable to find any eukaryotic homologs of SBF or Whi5 outside of fungi - reference_id: PMID:27162172 supporting_text: While we did not find either of the fungal regulators (SBF and Whi5) outside of Fungi, we did find animal-like cell cycle regulators in Archaeplastida, Amoebozoa, SAR, Haptophyta, Cryptophyta, Excavata and Metazoa blocks: - gene: cec-10 protein: UniProtKB:Q9N4L7 organism: Caenorhabditis elegans evidence: IBA source_node: PANTHER:PTN000917496 - protein: UniProtKB:A7S410 organism: Nematostella vectensis evidence: IBA source_node: PANTHER:PTN000917496 - description: 41 TreeGrafter IEAs on animal Swi6/Cdc10-family ankyrin proteins (corals, sea anemones, molluscs, annelids, a brachiopod, a priapulid, a beetle) and one plant (Carpinus fangiana), from PTN000917496, PTN001541507, PTN002887935, PTN000917511 and PTN002466252 would_wrongly_block: [] notes: Counts are exact-term non-NOT annotations in QuickGO on 2026-11-05 (4,809 total, 43 outside Fungi). The constraint fixes the symptom; the cause is PAINT node PTN000917496, whose MBF IBDs sit above metazoan members of PTHR43828 that lack the KilA-N/APSES DNA-binding domain. An IRD on the metazoan branch (or moving the IBD to the fungal subfamily) is the matching PAINT change. in_repo_agreement: - genes/worm/cec-10/cec-10-ai-review.yaml (REMOVE; propagation_review LINEAGE_OR_TAXON_MISMATCH) status: PROPOSED source: projects/RAR_MISANNOTATION.md (Wei et al. 2020, PMID:32470107, Table S3) - id: TPV-005 go_term: id: GO:0033309 label: SBF transcription complex constraint: only_in_taxon taxon: id: NCBITaxon:4751 label: Fungi existing_go_constraint: - constraint: only_in_taxon taxon: id: NCBITaxon:2759 label: Eukaryota - constraint: never_in_taxon taxon: id: NCBITaxon:4895 label: Schizosaccharomyces evidence: - reference_id: PMID:27162172 supporting_text: We were unable to find any eukaryotic homologs of SBF or Whi5 outside of fungi blocks: - description: 41 TreeGrafter IEAs, the same animal and plant Swi6/Cdc10-family proteins and PANTHER nodes as TPV-004 would_wrongly_block: [] notes: 5,482 exact-term annotations in QuickGO on 2026-11-05, 41 outside Fungi. Kept separate from TPV-004 because the two complexes are distinct terms with separate constraints today. cec-10 lost its SBF IBA after the 2019-06-03 release that Wei et al. analysed, but the TreeGrafter IEAs remain. in_repo_agreement: - genes/worm/cec-10/cec-10-ai-review.yaml (MBF terms removed on the same grounds) status: PROPOSED source: projects/RAR_MISANNOTATION.md (Wei et al. 2020, PMID:32470107, Table S3) - id: TPV-006 go_term: id: GO:0044548 label: S100 protein binding constraint: only_in_taxon taxon: id: NCBITaxon:7711 label: Chordata existing_go_constraint: - constraint: never_in_taxon taxon: id: NCBITaxon:4751 label: Fungi evidence: - reference_id: PMID:27764152 supporting_text: All S100 hits were within vertebrates, with the exception of four hits from the tunicate Oikopleura dioica. - reference_id: PMID:27764152 supporting_text: The presence of S100 proteins in tunicates and vertebrates (Olfactores), but not other chordates, suggests that the first S100 arose in the last common ancestor of tunicates and vertebrates - reference_id: PMID:18275604 supporting_text: There is no evidence for S100 proteins in invertebrates blocks: - gene: CG3226 protein: UniProtKB:Q9W3Y3 organism: Drosophila melanogaster evidence: IEA (InterPro2GO) source_id: InterPro:IPR037893 - description: 1,194 InterPro2GO IEAs from IPR037893 (Calcyclin-binding protein, CS domain) on CacyBP/SIP orthologs in plants, invertebrates and apicomplexans would_wrongly_block: [] notes: 5,596 exact-term annotations in QuickGO on 2026-11-05, 1,194 outside Chordata, all InterPro2GO. The S100 family is Olfactores-specific (tunicates plus vertebrates); Chordata is the conservative choice, and the tighter clade would block nothing more (no cephalochordate rows are affected). The matching mapping change keeps IPR037893 -> GO:0044548 but narrows it to chordates, as a broadMatch row in projects/INTERPRO/interpro2go.sssom.yaml, replaced by curated rule CR-0005 (rules/curated/). in_repo_agreement: - genes/DROME/CG3226/CG3226-ai-review.yaml (REMOVE) status: PROPOSED source: projects/RAR_MISANNOTATION.md (Wei et al. 2020, PMID:32470107, Table S4) - id: TPV-007 go_term: id: GO:0021987 label: cerebral cortex development constraint: never_in_taxon taxon: id: NCBITaxon:8782 label: Aves existing_go_constraint: - constraint: only_in_taxon taxon: id: NCBITaxon:33213 label: Bilateria evidence: - reference_id: PMID:23027930 supporting_text: Our findings establish that the layer 4 input and the layer 5 output cell types are conserved across the amniotes, but are organized into very different architectures, forming nuclei in birds, cortical areas in reptiles, and cortical layers in mammals. - reference_id: PMID:26154980 supporting_text: the six-layered neocortex structure could hardly be more dissimilar in appearance from the nuclear organization that is so conspicuous in the dorsal telencephalon of birds and other reptiles blocks: - gene: BTBD3 protein: UniProtKB:F1NI11 organism: Gallus gallus evidence: IBA source_node: PANTHER:PTN002908884 - gene: KIF14 protein: UniProtKB:A0A8V0YLM5 organism: Gallus gallus evidence: IBA source_node: PANTHER:PTN002790778 - gene: NTRK2 protein: UniProtKB:Q91987 organism: Gallus gallus evidence: ISS source_id: UniProtKB:P15209 - description: 1,756 IEAs on bird proteins (1,521 TreeGrafter, 212 Ensembl Compara, 23 InterPro2GO) would_wrongly_block: [] notes: 9,287 exact-term annotations in QuickGO on 2026-11-05, 1,759 in Aves; the 24 is_a/part_of descendants of GO:0021987 would also be excluded in birds. GO defines the cerebral cortex as "the outer layered region of the telencephalon"; the avian dorsal pallium is nuclear, not layered. Reptiles have cortical areas, so the constraint is never_in_taxon Aves rather than only_in_taxon Mammalia. Hippocampus development (GO:0021766) is not under GO:0021987 and is unaffected. GO:0021987 is part_of GO:0021543 pallium development, which stays available for birds, and the homology of avian pallial regions to neocortex is debated (PMID:39946453 argues for convergence), so this is the least certain of the four RAR proposals. The strongest argument against a hard constraint is the main clause of the PMID:23027930 quote, that the layer 4 and layer 5 cell types are conserved across amniotes, so only the layered architecture, not the cell types, is absent from birds. Uberon (checked 2026-11-06) puts no only_in or never_in constraint on UBERON:0000956 cerebral cortex; it asserts present_in_taxon NCBITaxon:117569 (Hyperoartia), with a note on convergent lamination in hagfishes, so Uberon also treats the cortex by its layered structure. conservative_alternative: constraint: none rationale: Leave GO:0021987 unconstrained and fix the bird annotations at their sources instead, moving them to GO:0021543 pallium development (the PAINT node PTN002908884, the TreeGrafter and Ensembl Compara rows). This avoids GO taking a position on avian pallial homology, at the cost of no automatic block on future bird annotations. in_repo_agreement: - genes/CHICK/BTBD3/BTBD3-ai-review.yaml (MODIFY to GO:0021543 pallium development) status: PROPOSED source: projects/RAR_MISANNOTATION.md (Wei et al. 2020, PMID:32470107, Table S4) - id: TPV-008 go_term: id: GO:0005118 label: sevenless binding constraint: only_in_taxon taxon: id: NCBITaxon:6656 label: Arthropoda conservative_alternative: constraint: only_in_taxon taxon: id: NCBITaxon:33208 label: Metazoa rationale: Metazoa blocks only the 13 rows that cannot be right under any reading, because receptor tyrosine kinases are absent outside animals. It is the fallback if GO reads the term as covering Sevenless orthologs, so that a ligand of vertebrate ROS1, the Sevenless ortholog, such as NELL2 could be annotated to it. No such annotation exists today (QuickGO, 2026-11-07). existing_go_constraint: [] evidence: - reference_id: PMID:41698918 supporting_text: compared to the invertebrate ortholog Sevenless (dROS1) - reference_id: PMID:41698918 supporting_text: revealing how trimeric NELL2 induces both receptor clustering and a conformational switch - reference_id: PMID:22550341 supporting_text: the receptor tyrosine kinases diversified extensively in each of the filasterean, choanoflagellate, and metazoan clades blocks: - description: 172 vertebrate InterPro2GO IEAs from IPR002956, 170 of them on GPRC5-family receptors (IPR051753) such as mouse Gprc5c, which the mouse Gprc5c review removes - description: 10 InterPro2GO IEAs from IPR002956 on non-arthropod invertebrates - description: 11 bacterial InterPro2GO IEAs from IPR002956, all transporters, permeases or uncharacterized membrane proteins (A0A1C3XK99, A0A1U9QRC0, A0A431TYL3, A0A4P6WYA3, A0A511NJ43, A0ABP6DAT4, A0ABP7TSN2, A0ABQ0BEM5, A0ABT8MF08, A0ABV9KKI3, A0ABZ2A5I0) - protein: UniProtKB:A0A0C3Q703 organism: Tulasnella calospora (fungus) evidence: IEA (InterPro2GO) source_id: InterPro:IPR002956 - protein: UniProtKB:A0ACF6VDB5 organism: Bienertia sinuspersici (plant) evidence: IEA (InterPro2GO) source_id: InterPro:IPR002956 would_wrongly_block: [] notes: >- GO:0005118 has no taxon constraint today (QuickGO, checked 2026-11-07). Its definition is "Binding to a sevenless (sev) protein, a receptor tyrosine kinase" (QuickGO, 2026-11-07); it names the sev gene product, and GO already restricts the pathway it feeds, GO:0045500 sevenless signaling pathway, to Arthropoda, so the binding term is given the same scope. All 440 exact-term annotations in QuickGO on 2026-11-07 are InterPro2GO IEAs from IPR002956 (Bride of sevenless, PRINTS PR01223); 245 are arthropod and 195 are not (172 vertebrate, 170 of them GPRC5-family; 10 other animals; 11 bacteria; 1 fungus; 1 plant), and this constraint blocks all 195. No ROS1 or vertebrate ROS1 ligand (P08922, Q99435) carries the term, so nothing correct is blocked. The mapping narrowing (projects/INTERPRO/interpro2go.sssom.yaml) and curated rule CR-0006 (rules/curated/), which requires IPR050726 and Diptera, fix the same rows from the annotation side. The first draft (2026-11-07) proposed Metazoa; the PR #6202 review argued for consistency with GO:0045500, and Metazoa is now the conservative alternative. Receptor tyrosine kinases arose separately in filastereans, choanoflagellates and animals, so neither Sevenless nor ROS1 has a homolog outside Metazoa. in_repo_agreement: - genes/mouse/Gprc5c/Gprc5c-ai-review.yaml (REMOVE on GO:0005118) - genes/DROME/boss/boss-ai-review.yaml (ACCEPT on GO:0005118) - rules/curated/CR-0006.yaml (sevenless binding only with IPR050726 AND Diptera) status: PROPOSED source: projects/RAR_MISANNOTATION.md (follow-up to the IPR002956 InterPro2GO review)