Skip to content

Copper Metabolism Disorders — GO Pathway Annotation Review (2026-08-28)

Review of every KB entry involving disorders of copper metabolism, focused on how Gene Ontology terms are used to describe the disease pathways. This document records both the review and the curation pass that acted on it: every finding below was fixed in the same-dated pass, and the "Fix applied" notes and the Validation section at the end record what changed.

Scope

The curated grouping kb/groupings/Disorders_of_Copper_Metabolism.yaml (closeMatch to MONDO:0017762) defines the core set, and its membership logic was confirmed correct against the entries:

Entry Gene Direction of copper defect
Wilsons_Disease ATP7B (hgnc:870) Overload — failed biliary excretion
Menkes_Disease ATP7A (hgnc:869) Deficiency — failed intestinal export
MEDNIK_syndrome AP1S1 (hgnc:559) Mixed — AP-1 trafficking of both pumps
Huppke-Brendel_syndrome SLC33A1 (hgnc:95) Low serum copper/ceruloplasmin, secondary

Also reviewed because they touch copper biology:

  • aceruloplasminemia (CP, hgnc:2295) — deliberately excluded from the grouping as iron-primary; the exclusion rationale in the grouping is sound and the entry's annotation bears it out (see below).
  • The mitochondrial copper-chaperone trio SCO1-Related_COX_Deficiency, SCO2-Related_Fatal_Infantile_Cardioencephalomyopathy, COX11-Related_COX_Deficiency — copper delivery to cytochrome c oxidase, grouped under Mitochondrial_Complex_IV_Deficiency and conforming to the complex_iv_assembly_deficiency module.
  • com_Wilsons_Disease__Osteoporosis (comorbidity).

All gene CURIEs, MONDO anchors, and every GO binding's id/label pair in these files were verified against the authority-backed caches (cache/go/terms.csv, cache/hgnc/terms.csv, enum membership caches). No hallucinated or mislabeled GO term was found — every bound label matches the canonical label, including recent GO additions.

What is done well

  1. Canonical labels and modern terms. The set uses current GO labels throughout, including terms many curation pipelines miss: GO:0160119 cuproptosis and GO:0097707 ferroptosis as first-class cell-death nodes in Wilson disease, GO:0150076 neuroinflammatory response, GO:0034102 erythrocyte clearance, and GO:0006879 under its current label intracellular iron ion homeostasis (aceruloplasminemia).

  2. Chemically precise molecular function. Wilson disease binds GO:0140581 P-type monovalent copper transporter activity — correctly Cu(I), not a generic copper-transporter term — on the ATP7B nodes.

  3. Module conformance carries shared GO vocabulary. Wilson's fibrosis cascade conforms to fibrotic_response (TGF-β GO:0007179 INCREASED, ECM organization GO:0030198 INCREASED, collagen biosynthesis GO:0032964 INCREASED) and its hemolysis arm to hemolytic_anemia_erythrocyte_destruction. The SCO1/SCO2/COX11 trio conforms to complex_iv_assembly_deficiency, and their use of the general GO:0006825 copper ion transport for chaperone-mediated copper delivery is the module's own contract — appropriately general, since metallochaperone hand-off is not transmembrane transport.

  4. Secondary copper phenotypes are not force-fitted to copper GO terms. Huppke-Brendel is modeled through its actual proximal pathway — acetyl-CoA transport (GO:0035348/GO:0008521, ER GO:0005783) → protein acetylation (GO:0006473) → protein secretion (GO:0009306 DECREASED) → low ceruloplasmin/copper — rather than borrowing copper-transport terms the gene does not participate in. MEDNIK similarly leads with AP-1 trafficking terms (GO:0006886, GO:0016192) and only introduces copper GO terms on the downstream copper-pump-trafficking and hepatic-overload nodes. This is the right application of the "no term beats a bad term" rule.

  5. Aceruloplasminemia annotates iron, not copper. Its three nodes bind GO:0004322 ferroxidase activity (DECREASED), GO:0006826 iron ion transport, and GO:0006879 — consistent with the grouping's decision that its proximal lesion is iron export despite the ceruloplasmin connection.

Findings and fixes applied

Every finding below was fixed in the same-dated curation pass (commit on claude/copper-metabolism-disorders-w3e4il); the "Fix applied" note under each records what changed. All eight edited entries pass just validate-disorders (schema + terms + reference snippets), the duplicate-key and entity-ref checks, and the offline title/grading/length/hyphen/environmental gates. Candidate GO terms were verified against OLS before binding, and matching history records were added under history/disorders/.

Ranked by impact on the machine-readable pathway description. None was a validation failure; all files passed the term contract before and after.

1. Wilson disease GO descriptors carry no direction modifiers

Every copper-pathway GO binding in Wilsons_Disease.yaml lacks a modifier, so the direction of the lesion lives only in node names and prose ("Impaired Biliary Copper Excretion", "Hepatic Copper Accumulation"). The sibling entries annotate direction (Menkes GO:0006825 DECREASED on brain copper transport; MEDNIK ABNORMAL; SCO1/SCO2/COX11 DECREASED throughout), so Wilson is the outlier in the grouping, and pathograph/KGX consumers cannot see that copper transport is decreased in the very entry that is the copper-overload archetype. Suggested minimal fix: DECREASED on GO:0006825 and GO:0140581 for the "ATP7B Copper-Trafficking Defect", "Impaired Biliary Copper Excretion", and "Impaired Ceruloplasmin Loading" nodes. Menkes' apex node ("ATP7A-mediated copper export failure", GO:0006825 unmodified) has the same gap even though its child nodes are modified.

Fix applied. DECREASED added to GO:0006825 and GO:0140581 on the three Wilson ATP7B nodes and to the Menkes apex node (which also gained the GO:0140581 copper-transporter MF, DECREASED). Downstream oxidative-stress and homeostasis nodes across both entries received INCREASED/ABNORMAL modifiers to match.

2. One GO triple copy-pasted across mechanistically distinct nodes

Wilson's "ATP7B Copper-Trafficking Defect" and "Impaired Biliary Copper Excretion" carry the identical BP/MF/CC triple (GO:0006825 + GO:0140581 + GO:0005794 Golgi apparatus), and "Impaired Ceruloplasmin Loading" reuses two of the three. In GO space the three nodes are near-indistinguishable, while their own descriptions distinguish them precisely: the biliary-excretion node describes ATP7B trafficking to the bile canalicular membrane, for which GO:0035434 copper ion transmembrane transport (already in the cache and used by Cardiomyopathy-Hypotonia-Lactic_Acidosis_Syndrome) is the more specific accurate BP, and the Golgi CC is arguably wrong — the site of that step is the apical/canalicular membrane, not the TGN. Aceruloplasminemia has the same pattern in miniature: the identical iron BP pair is stamped onto all three of its nodes, with modifier: ABNORMAL uniformly, so the ferroxidase node and the two accumulation nodes differ only by the MF binding.

Fix applied. The Wilson biliary-excretion node now binds GO:0035434 copper ion transmembrane transport (DECREASED) with GO:0016324 apical plasma membrane as the canalicular CC, replacing the Golgi triple. The aceruloplasminemia nodes were left on the shared iron pair but are now differentiated by biological_scale (MOLECULAR / ORGANISM / TISSUE); their ABNORMAL homeostasis modifier is correct under the finding-5 convention, with iron level carried by the INCREASED iron chemical_entities.

3. The treatment-targeted hub nodes have no ontology grounding at all

Wilson's "Hepatic Copper Accumulation" and "Systemic Copper Distribution" are the two most connected nodes in the KB's copper pathograph — every chelator, zinc, and dietary treatment target_mechanisms link points at them — yet neither carries any GO (or other ontology) binding. Candidates worth validating through the dismech-terms workflow: GO:0055070 (or the more specific GO:0006878 intracellular copper ion homeostasis, not yet in the cache) with an INCREASED/ABNORMAL modifier, and GO:0046688 response to copper ion for the downstream injury claims. As curated, an export of these nodes is free text only.

Fix applied. "Hepatic Copper Accumulation" now binds GO:0006878 intracellular copper ion homeostasis (ABNORMAL) and "Systemic Copper Distribution" binds GO:0055070 copper ion homeostasis (ABNORMAL).

4. Cuproenzymes are named in prose but their molecular functions are never bound

A cross-entry pattern: the copper-deficiency arm of the group repeatedly names specific cuproenzymes — dopamine-β-hydroxylase, cytochrome c oxidase, lysyl oxidase, tyrosinase — in description text, but the only cuproenzyme MF bound anywhere in the group is GO:0004322 ferroxidase activity (in aceruloplasminemia). Concretely:

  • Menkes "Lysyl oxidase deficiency and connective tissue fragility" binds only the consequence (GO:0030198 ECM organization) and not the cause; a DECREASED protein-lysine 6-oxidase activity MF binding would make the node's own name machine-readable.
  • Menkes "Cuproenzyme deficiency and neurodevelopmental injury" binds the processes (GO:0042423 catecholamine biosynthesis, GO:0006119 oxidative phosphorylation) but not dopamine β-monooxygenase or cytochrome-c oxidase activity, despite the description and the DOPA:DHPG biomarker resting on DBH specifically.
  • Menkes "Hair-shaft keratinization and pigmentation enzyme dysfunction" has no ontology bindings at all (tyrosinase/melanin biosynthesis are the obvious candidates).
  • Wilson "Impaired Ceruloplasmin Loading" could mirror aceruloplasminemia's GO:0004322 DECREASED — the two diseases converge on lost ceruloplasmin ferroxidase activity by different routes, and a shared MF binding would make that convergence queryable.

Fix applied. All four cuproenzyme MFs were verified against OLS and bound (DECREASED): Menkes lysyl oxidase GO:0004720, dopamine β-monooxygenase GO:0004500, cytochrome-c oxidase GO:0004129, and tyrosinase GO:0004503 (with melanin biosynthesis GO:0042438 on the hair node); Wilson's ceruloplasmin-loading node gained the shared GO:0004322 ferroxidase activity binding, making its convergence with aceruloplasminemia queryable. New mechanistic citations were added for the previously prose-only claims: PMID:32381719 (Elesclomol/COX in Menkes) and PMID:18650808 (ATP7A→tyrosinase in melanosomes).

5. Inconsistent modifier semantics on homeostasis terms

Menkes "Systemic copper deficiency" annotates GO:0055070 copper ion homeostasis with DECREASED; MEDNIK "Hepatic copper overload" annotates the same term with ABNORMAL. The two nodes describe opposite derangements, but the annotations don't encode that — and "decreased homeostasis" conflates the copper level with the homeostatic process (if anything, a deficiency state means homeostasis has failed, not diminished). A convention worth adopting KB-wide: homeostasis-class terms take ABNORMAL/DYSREGULATED, and directionality is carried by the transport/level-specific binding or the node itself.

Fix applied. The convention was adopted across the copper set: Menkes "Systemic copper deficiency" changed from DECREASED to ABNORMAL on GO:0055070, and the Wilson and MEDNIK homeostasis nodes were made consistent. Direction now rides on the copper-transport MF/BP bindings and the copper chemical_entities level, not on the homeostasis process term.

6. Menkes OXPHOS binding sits in mild tension with its own cited evidence

The "Cuproenzyme deficiency" node binds GO:0006119 oxidative phosphorylation (unmodified) with PMID:27226607 among its support — but the same paper, quoted on the adjacent "Mitochondrial redox imbalance" node, reports that redox misbalance "does not significantly affect … the activity of respiratory complex IV" in that model. The OXPHOS-impairment claim is defensible from the broader cuproenzyme literature, but the specific 27226607 snippet on this node ("ATP7A activity protects mitochondria from excessive copper entry") supports the copper-accumulation redox node, not the deficiency-driven cuproenzyme node it is attached to. Worth re-homing that evidence item and deciding whether OXPHOS merits a DECREASED modifier backed by COX-specific human data.

Fix applied. The misplaced PMID:27226607 item was removed from the cuproenzyme node (it remains, correctly, on the redox node) and replaced with PMID:32381719, which attributes the Menkes mitochondrial energy deficit specifically to cytochrome c oxidase dysfunction. GO:0006119 oxidative phosphorylation now carries DECREASED, grounded on that COX-specific evidence plus the new GO:0004129 cytochrome-c-oxidase MF.

7. Secondary observations (non-GO)

  • Cell-death edge direction. Wilson previously modeled Hepatocyte Injury → Cuproptosis and → Ferroptosis (death programs as leaf consequences), while the entry's own hypothesis files treat cuproptosis as a mechanism of hepatocyte injury. Fixed: the two death nodes are now wired Hepatic Copper Accumulation → Cuproptosis → Hepatocyte Injury (with PMID:35298263 grounding copper's direct binding to lipoylated TCA proteins) and Ferroptosis → Hepatocyte Injury (with PMID:41966025), so the graph and the hypothesis register now agree. The former leaf edges were removed.
  • ICIMD classification was inconsistent across the grouping. Fixed: Menkes and MEDNIK gained classifications.icimd_category: copper_metabolism; aceruloplasminemia gained an iron_metabolism classification recording why it is iron-primary. Huppke-Brendel was deliberately left unclassified on the ICIMD axis — its proximal lesion is the acetyl-CoA transporter and its low copper is explicitly secondary, so a copper_metabolism tag would contradict the entry's own framing; its placement between the acetylation/CDG and copper groups is a genuine curator decision, not an omission to paper over. This means the grouping's rationale sentence claiming all four fall in the ICIMD Copper group is itself slightly overstated for Huppke-Brendel and is worth softening.
  • biological_scale is now tagged on every node in all eight entries (MOLECULAR pump/enzyme defect → CELLULAR → TISSUE/ORGANISM injury).
  • Wilson's deprecated prevalence.percentage fields were removed; the verbatim source phrasing was preserved in notes.
  • The comorbidity com_Wilsons_Disease__Osteoporosis binds GO:0046849 bone remodeling correctly (label verified); left unchanged.

Literature added

  • PMID:35298263 (Tsvetkov 2022, Science) — copper binding to lipoylated TCA proteins; grounds the rewired Wilson cuproptosis edge.
  • PMID:32381719 (Guthrie 2020, Science) — Elesclomol/COX rescue in a Menkes model; grounds the Menkes cytochrome-c-oxidase and OXPHOS annotations.
  • PMID:18650808 (Setty 2008, Nature) — ATP7A sustains melanosomal tyrosinase; grounds the Menkes hair-node tyrosinase/melanin bindings.
  • PMID:42018271 (2026) — January 2026 FDA first approval of copper histidinate (ZYCUBO) for pediatric Menkes disease, added to the treatment.
  • NCT04537377 (VTX-801) and NCT04884815 (UX701) — the two active Wilson disease AAV gene-therapy trials, added as clinical_trials.

Remaining follow-up (not in this pass)

  • Soften the Disorders_of_Copper_Metabolism grouping rationale's claim that all four members fall in the ICIMD "Copper" group, given Huppke-Brendel's acetyl-CoA-transporter primary lesion (see finding 7).
  • Earlier suggested items now done in this pass: direction modifiers (finding 1), node differentiation (finding 2), hub-node grounding and cuproenzyme MFs (findings 3–4), the homeostasis-modifier convention and the re-homed Menkes evidence item (findings 5–6), and the ICIMD/biological_scale backfill (finding 7).

Post-review refinements (2026-08-30)

Applied after the ai4c-reviewer pass on PR #9852, alongside a merge with a fast-moving main:

  • PMID:18650808 (Setty et al.) regraded MODEL_ORGANISM → IN_VITRO: the study is entirely cultured mouse melanocyte lines, and CLAUDE.md places cultured animal cells under IN_VITRO. The hypopigmented-mice line in that abstract is a citation to the paper's own refs, not its data.
  • The Menkes PMID:32381719 snippet was swapped from the paper's background sentence to its own in-vivo result (elesclomol raising brain cytochrome c oxidase), which grounds the COX/OXPHOS annotations from the intervention.
  • MEDNIK's GO:0070830 tight-junction modifier changed DECREASED → ABNORMAL (the source reports mislocalization and functional barrier loss, not reduced assembly), with the CaCo2-AP1S1-KO measurement added as its own IN_VITRO evidence item — matching the finding-5 homeostasis convention.
  • The Wilson biliary node kept GO:0035434 as its BP but regained the trans-Golgi GO:0005794 cellular component alongside the canalicular GO:0016324, since the node still describes resting TGN localization.
  • icimd_category: complex_iv_subunits_and_assembly_factors was added to SCO2 and COX11 to match the SCO1 sibling.
  • Both Wilson gene-therapy trials are now curated at their earliest registered phase (the single-valued phase slot cannot express a range).

The merge kept main's independent additions to these entries — the genes blocks on SCO1/COX11 and an INCREASED modifier on the Wilson cuproptosis node — alongside this branch's copper-delivery molecular functions and the cell-death rewiring.

Validation

All eight edited entries pass just validate-disorders (schema + terms + reference snippets), just check-duplicate-keys, just check-entity-refs, and the offline title/grading/length/hyphen/environmental gates. New GO terms were verified against OLS before binding, new reference titles match their cache frontmatter, and a history/disorders/ record was added per edited entry in both passes (just validate-history-all clean).