The single scaffold core reflects COG5-specific assembly evidence, rather than transfer of COG4's recruitment mechanism. COG5 cooperates with COG7 to build a COG6-containing subcomplex. The original experiments use human cDNAs in rabbit reticulocyte lysate and explicitly qualify binary directness pending purified-protein tests. The COG7-source binding assertion is refined to this structural activity; other positive pairs remain distinct. No NEW annotation is added. PMID:15047703
The historical GTC-90 experiments use bovine complex and CHO donor/acceptor Golgi, with rat microscopy. Antibody inhibition supports complex function but does not isolate COG5 tethering or fusion. The human cDNA work is a separate experimental component. The subsequent complex-characterization paper primarily perturbs COG1/COG2 and images COG1/COG7; these assays must not be reassigned to COG5. [PMID:9792665; PMID:11980916]
In the human knockout panel, COG5 loss destabilizes COG7 and reduces COG6. The Golgi retains dilated mini-stacks and is less severely affected than several other COG knockouts. The integrated toxin-transport readout is interpreted with its compartment and cell-panel limits. COG5 supplies assembly activity to trafficking machinery, rather than catalyzing the glycan reactions whose distribution changes. PMID:27066481
The first COG5-CDG abstract explicitly reports a patient fibroblast trafficking defect and wild-type cDNA rescue. Later full patient studies show loss of COG5/COG7 abundance and delayed GalT-GFP redistribution. These findings support transport and Golgi organization, without manufacturing neurological-development or glycosyltransferase activities. The human variant study distinguishes Golgi/stress assays from separate mouse retinal PERK experiments. [PMID:19690088; PMID:23228021; PMID:33277529]
The 2026 study provides positive cellular COG5–ATP7A association, protein-turnover and rescue data. It proposes an ATP7A recycling/localization mechanism; that trafficking step is not directly resolved. Copper accumulation and respiratory phenotypes therefore remain mechanistic context and a question for separation-of-function experiments, not a new COG5 copper-transporter or OXPHOS annotation. Its patient-cell and variant interpretation limits are retained. PMID:41824529
Golgi association and a cytosolic complex pool are compatible distributions. The nine Reactome location assertions resolve through Golgi COG5 entity 6808725 or TGN entity 6813869 to reference entity 52520/Q9UP83. Membership in a pathway event does not assign every Rab, SNARE, uncoating or fusion step to COG5.
All four original product objects remain exact. Official sequences independently reconstruct to 860 residues for displayed Q9UP83-2 (name 1), 839 for Q9UP83-1 (name 4), 823 for Q9UP83-3 (name 3), and 829 for Q9UP83-4 (name 2). The alterations respectively remove residues 594–614, 824–860, or 1–31 from the displayed sequence. Neither matching length nor an old unversioned construct accession establishes equal function of all four products.
The NK-like YTS proteomics source is abstract-only in the selected cache. Its publisher landing page was read, but the supplemental spreadsheet returned 403 and the full-text link redirected to the abstract. The exact COG5 table entry remains UNVERIFIED. The broad membrane assertion is retained as non-core because independent COG5 evidence establishes peripheral membrane association; this is not a claim that the YTS table or a plasma-membrane location was confirmed. PMID:19946888
The explicit campaign policy in projects/CLINGEN_MENDELIAN.md retains supported experimental generic-binding assertions as non-core unless an informative refinement is established. The four source pairs have eight exact 42-field IntAct records among 72 raw export records. Two reciprocal in-vitro records support each original assembly pair. BioPlex supplies HEK293T and HCT116 associations, with COG7 bait in one HEK293T record and COG5 bait in the HCT116 record; COG5 is bait for both COG4 records. AP-MS co-purification is not purified binary affinity. [PMID:15047703; PMID:33961781]
All 31 raw GOA source tuples match the 31 original assertions without deduplication, and all four products and 19 original reference identity pairs are preserved. Four genuine clinical/mechanistic references are added. Five current donor/target joins contain 19 records. Their literal database references are provenance, not claims that every donor paper was fetched or read. Current yeast, fly and human records support the retained biological assertions; the historical PAINT node/alignment remains uninspected and its source status stays UNRESOLVED. Target self-evidence is legitimate descendant grounding. The exact GO-CAM index has no target hit; absence is not used to infer missing function.
The default Falcon research attempt genuinely timed out and its configured fallback returned 401 insufficient quota. No provider report was invented. Normal fetches, original source bytes and all failed routes remain archived; manual reading and a bounded independent consultation support the assessment.
The selected 11980916 and 15047703 abstracts remain protected. A separate genuine normal full variant of 11980916 and authentic external original text for 15047703 and 9792665 were read within the recorded windows. The owner ledger distinguishes these reads from additional independent ROOT consultation reads. The BioPlex selected body contains Introduction/Discussion only, not full Methods/Results. Clinical 19690088 remains abstract-only; the other three added papers have genuine normal full text. The incidental 23186163 phosphoproteome abstract was fetched but is not cited or proposed for publication.
The publication index contains compact literal record projections, source URLs/hashes and access limits. It is authored provenance, not primary support. Raw original article text stays in TMP. Three short anchors occur once each in YAML and are not repeated in these notes or index. Existing main publication/Reactome/GO-CAM sources stay byte-exact. The normal full 27066481 source also exists in the independently published COG4 branch and will be checked for exact shared bytes before publication. No canonical file, Git state or existing history was changed by this owner.
Genuine normal validation, calculated-status checking and rendering passed. The four advisories are three supported generic-binding retentions under the explicit campaign policy and one GO:0017119 label warning: the saved current official term is “COG complex,” while the validator expects its older label “Golgi transport complex.” The source term and current core label are retained. Calculated status remains DRAFT. No validation skips, source changes or added annotations were used to suppress these advisories.
The COG7-source scaffold refinement is retained on performed cooperative assembly, rather than the strength of one binary contact. The original study reports a weak specific COG5–COG7 pair and shows that COG5 plus COG7 incorporates COG6 whereas either alone does not. The stronger COG4 pair remains a separate positive non-core assertion. The authors’ relative use of direct does not remove their explicit purified-protein caveat. No COG4-specific Sly1/STX5 mechanism is transferred.
The following public primary texts make the previously external method claims checkable. The source index retains exact response hashes and selected ranges; those hashes are provenance, not substitute evidence. The earlier protected abstract caches are unchanged.
Clinical trafficking support is now linked under the existing retrograde-recycling assessment: PMID:19690088 reports wild-type COG5 cDNA rescue, and PMID:23228021 provides the patient GalT-GFP/BFA time course. BFA-induced Golgi-to-ER redistribution is kept separate from intra-Golgi cargo flux. GO:0006890 was explicitly considered. Its definition names vesicle-mediated Golgi-to-ER transport, whereas the directly inspected clinical assay follows BFA-induced tubular redistribution. COG5’s structural assembly work is established, but a distinct route-specific vesicle step is not resolved here; no NEW process is proposed on these observations alone. This is not a denial of Golgi-to-ER participation or a requirement for autonomous transport chemistry.
The broad membrane annotation remains non-core because the current definition includes attached proteins and independent fractionation/localization supports peripheral COG5 association. The original YTS target-table relationship remains UNVERIFIED. These are separate judgments, and no plasma-membrane or integral-membrane claim follows.
The nine Reactome location assessments now identify their particular tethering, uncoating, SNARE, RAB or fusion context. They retain location through the represented COG complex; the event’s other biochemical steps are not automatically activities of COG5. GARP–COG joint capture remains explicitly unestablished in the relevant event.
All original source objects, products, annotation actions, core terms and reference identities remain intact. Historical notes are retained as provenance; this appendix records the current assessment. Three new short excerpts above occur once each and are counted with existing YAML quotations. Supported generic interactions remain non-core under the explicit campaign policy in projects/CLINGEN_MENDELIAN.md; no interaction is discarded merely for low specificity.
Ordinary validation, status dry-run and rendering passed on the final YAML with four advisories: three supported generic-binding non-core rows under the explicit campaign override, and the existing GO:0017119 COG-complex label difference against the validator’s legacy label. No schema error or status mismatch occurred. The exact original source labels are preserved. The final notes/index context was rendered again after this validation record was appended. The isolated validator’s ordinary term-cache and lockfile changes were recorded separately; no validation source was patched.