One SNARE-binding core integrates COG6's recruitment and assembly contribution at Golgi membranes. Human COG6 associates with SNAP29 in yeast two-hybrid and HeLa co-immunoprecipitation assays. Purified binding is demonstrated separately for STX6 and GS27; this directness is not automatically assigned to SNAP29. N-terminal mapping and selective SNARE recruitment after artificial mitochondrial targeting establish more than a generic loss phenotype, while the artificial location is not a native mitochondrial function. COG6 recognizes SNAREs rather than performing their helical-bundle zippering. The two existing SNAP29 assertions are refined to GO:0000149; no redundant NEW annotation is added. [PMID:21807881; PMID:23057818]
The original COG assembly study used human cDNAs in rabbit reticulocyte lysate. COG5 plus COG7 cooperatively incorporated COG6, and omission of COG6 affected COG8 assembly. This establishes structural participation but does not isolate a COG6 scaffold interface or justify transferring COG4's distinct Sly1/STX5 mechanism. The study explicitly treats binary directness as relative pending purified-protein confirmation. Its selected normal abstract stays exact; the independent bounded consultation read the authentic external original Methods and Results. PMID:15047703
Patient fibroblasts carrying a COG6 defect have altered complex fractionation and delayed brefeldin-A-induced Golgi-to-ER redistribution, rescued by wild-type COG6. The selected cache is abstract-only; detailed fractionation, complementation and transduction Methods were actually read from indexed author-uploaded original full text. The institutional download returned 403 HTML and is not treated as a PDF. BFA redistribution is a cellular retrograde readout, not a purified intra-Golgi transport assay. PMID:20605848
Human COG knockout cells directly establish COG6-dependent Golgi architecture and lobe-B stability. COG6-deficient Golgi cisternae are severely altered. Integrated transport and glycosylation assays support trafficking machinery function without assigning glycosyltransferase chemistry to COG6. The toxin experiment's representative cell panel is kept separate from directly inspected COG6 findings. COG6 is an exception to the immature cathepsin-D secretion phenotype in the tested panel; no universal secretion defect is inferred. PMID:27066481
The Shaheen syndrome abstract reports low COG6/STX6 expression despite normal transferrin glycosylation in the described family. This limits a universal glycosylation phenotype; it does not negate COG6's trafficking role or establish additional neurological-process annotations. No NEW process is proposed. PMID:23606727
All three original alternative-product objects remain unchanged. Official sequences reconstruct to 657, 615 and 75 residues. The short product ends before the experimentally mapped 76–150 SNARE-binding region and carries the official possible nonsense-mediated-decay caveat. The longer alternative sequence is not declared equivalent to the tested full-length construct solely from retained domain boundaries.
All 68 original binding assertions have exact paper–COG6–partner matches in 193 complete 42-field IntAct/MINT records. BFG-Y2H, HuRI and the viral screen are distinguished from BioPlex co-purification. Human mapped sequence identity does not turn yeast expression into a human-cell assay. Global orthogonal validation of a screen is not assigned to every target pair. Partner-class refinements describe the tested partner and do not import its catalytic, transcriptional, transport or cytoskeletal activity into COG6.
The campaign policy in projects/CLINGEN_MENDELIAN.md retains experimentally supported, otherwise uninformative binding as non-core. Specific class refinements use official definitions and actual product scope. NEK6-4 residues 26–338 match the complete 313-residue canonical kinase sequence. DTNB-5 retains 560 residues, but its exact cytoskeletal-component class is unresolved. UFSP1's current 218-residue sequence matches the mapped ROGID; the historical clone's initiation site is unresolved after the initiation correction, so the screen is not described as having used the hypothetical 142-residue form. These positive pairs remain represented. The viral nsp14 and nsp16 assertions concern processed proteins, not enzyme activity of an undifferentiated polyprotein. [PMID:32296183; PMID:36217029]
The BioPlex selected source body has Introduction and Discussion, not full Methods/Results. Actual pair provenance comes from complete literal database records. BFG-Y2H selected Methods and Results and HuRI Results/Discussion with assay-validation legends were read within the recorded scope. [PMID:27107012; PMID:28514442; PMID:33961781]
Six current donor/target joins contain 19 exact same-term records. Human target self-evidence is legitimate descendant grounding. The extant records support the retained biological judgments; the historical PTN000487537 tree, alignment and IBD placement were not inspected, so its source status remains UNRESOLVED. Donor number is not a confidence proxy. Literal donor references in the index are database provenance, not additional claims of full-paper access.
The nine Reactome assertions resolve through COG complex entities to COG6 reference entity 52522/Q9Y2V7. Golgi and trans-Golgi network compartments remain source-specific. Event membership does not make COG6 the Rab enzyme, coat-removal factor or SNARE fusion catalyst. The RAB6 event explicitly allows separate GARP and COG populations rather than demonstrating a single combined tether. The selected GO-CAM index has no target entry; absence does not establish a missing function.
All 89 original source objects, all three products and all 20 original reference identities are preserved. Three genuine supplemental references are added. Current GOA assertions reconcile exactly, and no NEW annotation is manufactured. Existing main publication, Reactome and GO-CAM bytes are protected; fresh normal variants remain archived. The PTHR21506 family files are genuinely fetched additions absent from the checked main, not pre-existing protected files. Shared 27066481 content is compared with current COG4 and COG5 publication heads at the final source gate.
The default Falcon attempt genuinely timed out; its configured fallback returned 401 insufficient quota. No provider report was invented. Manual evidence, genuine normal fetches and two bounded independent consultations supply the research record. The owner read ledger separates actual selected paragraphs, complete available abstracts and indexed original windows from the consultations' additional reads. The very long concatenated BFG Methods paragraph was truncated and is not claimed as completely read; smaller complete Methods entries are recorded instead.
The compact source index is provenance, not primary support. Raw full text, complete indexed responses and literal record arrays remain in TMP. Three short primary anchors appear once each in YAML and are not repeated in notes or index. All source URLs, selected hashes and declared access limits remain checkable. No canonical files, Git state or existing history were altered by this owner.
Genuine isolated normal1 validation, calculated-status checking and rendering passed. The 44 advisories comprise 43 deliberate generic-binding retentions under the explicit campaign policy and one current GO:0017119 label versus the validator's older label. The saved official response names COG complex; no original machine term is rewritten to satisfy an older label cache. Calculated status remains DRAFT. The final notes correct only the absent-main family-file disposition and append this validation record; YAML science and source-index bytes are unchanged. The final ordinary render includes these notes.
After normal1 and the preserved v1 seal, the independent reviewer identified overly restrictive class wording in nine non-core interaction reasons. GO:0008092 requires a protein component of the cytoskeleton, not a separate mechanical or structural criterion. Rows 20, 35, 42, 44, 47, 55, 68, 75 and 76 now use that actual membership boundary. All actions, source objects, products, reference identities, core terms and source-index bytes remain exact. The nine-leaf parsed delta and complete v1 snapshot are preserved. Genuine normal2 validation, status checking and rendering pass with the same 44 explained advisories; calculated DRAFT remains consistent.
The earlier no-NEW conclusion is superseded only for GO:0042147. COG6 recognizes STX6 and supports its SNARE assembly while cellular cargo experiments test endosome-to-TGN retrieval. This establishes performed machinery work, not simply dependence of a cargo on COG6. One NEW IMP annotation and the corresponding process in the existing SNARE core now capture that route. Intra-Golgi source rows remain intact; their reasons and citation lists distinguish intra-Golgi context from endosome-to-TGN and Golgi-to-ER assays. No isolated COG6 fusion or universal isoform activity is claimed. [PMID:21807881; PMID:23057818]
The exact current comparator query found no GO:0042147 records for COG3/COG4, and four each for the analogous GARP subunits VPS53/VPS54. VPS53 has involved_in IMP; VPS54's IMP is acts_upstream_of_or_within, with other involved_in records. The mixed result does not establish a role-based exclusion convention. Typed is_a/part_of ancestry and descendants do not overlap the existing/replacement/core terms. Untyped graph traversal includes other relations and is not the redundancy test. The authenticated current GO-CAM index has no COG6 entry. No comparator primary-paper reread is claimed.
RSPH14 and SAXO4 refinements retain their specific positive human screen pairs and explicit ortholog-based component classes. Official mouse Q9D3W1 radial-spoke and Q148A4 microtubule-inner-protein records carry experimental component evidence. Those database records were read; their cited donor papers were not independently recovered in this amendment. Exact human product mapping and SAXO4-2's internal deletion remain distinct from a direct human component assay. Inference depth alone is not negative evidence.
The original Rab study is now included with the species and construct limits in its reference assessment; the two existing human RAB6B screen refinements remain unchanged. A separate Rab core is deferred because native human isoform-specific recognition and its mechanistic separation remain unresolved. PMID:18256213
STX5 is explicitly named alongside STX6, GS27 and SNAP29 in the biological core description. Its directed Y2H and cellular association are not assigned the purified directness established for STX6/GS27. Assay and construct details remain in source assessments rather than dominating the core description. PMID:23057818
The original notes prefix is retained as session provenance. One additional short primary anchor is in the NEW row and is not duplicated here. The reachable source-index reference documents provenance only; it is not cited as primary biological support. All 89 original source objects, their actions, three products, 23 reference identities, protected caches and existing history remain unchanged. One genuinely fetched PMID cache is appended; no provider output was manufactured.
Genuine normal2 validation, status checking and rendering all passed on the final science YAML. The 44 advisories remain 43 supported generic-binding retentions under the explicit campaign policy and the existing GO:0017119 current-label versus validator-label difference. Status remains DRAFT. Normal1 stopped because the isolated copy omitted script executable modes; restoring the authenticated source modes changed no tooling content, and its failed receipt is preserved. The ordinary uv invocation recalculated the isolated workspace lock without adding or changing registry package versions and added one genuine GO term-cache row; neither derived file is proposed for publication.
The final finite guard authenticated the unchanged PR head and complete feedback, 34 original preimages including one history, and the new cache's absence from main and other open PRs. At main 905a35aeb3efd4a8b4c14248748f25ba8a7e358f, complete script-tree comparison showed only the unrelated auto_merge_ready_prs.py helper changed; all invoked validation/status/render tooling and source preimages remained exact. The 340 complete open-PR inventories reuse unchanged authenticated head/base lists and freshly retrieve changed inventories. ROOT repeats application checks immediately before publication.