COG8

Summary: COG8 associates with Golgi membranes.

Reason: The Swiss-Prot Golgi-location mapping agrees with human COG8 localization and its peripheral association with the COG tethering complex. This does not imply a transmembrane segment.

Summary: The human knockout study places COG8 in the Golgi-associated COG complex.

Reason: The source studies the complete set of human HEK293T COG knockouts and directly includes COG8. Its Golgi-complex context and independent human localization support the NAS membrane assertion; individual phenotypes are not all treated as direct localization assays.

Summary: COG8 is represented at the Golgi membrane in cis-Golgi vesicle tethering.

Reason: The model places the COG complex with USO1 and Golgin tethers during cis-Golgi vesicle capture. Complex membership supports the original Golgi membrane location, without assigning each modeled tether contact to COG8.

Summary: COG8 is represented at the Golgi membrane in COPI vesicle uncoating.

Reason: The modeled COPI uncoating step has an explicitly unresolved mechanism; coat release and later NSF-dependent SNARE disassembly do not establish COG8 uncoating or ATPase activity. COG8 is retained as a membrane-associated COG constituent in the event.

Summary: COG8 is represented at the Golgi membrane in cis-Golgi SNARE engagement.

Reason: The event models YKT6 engagement with cis-Golgi t-SNARE combinations during fusion. The COG8 entity chain supports membrane context, not that COG8 is a SNARE, performs zippering or has a directly measured YKT6 interface.

Summary: COG8 is represented at the Golgi membrane in Golgi SNARE recognition by the COG complex.

Reason: The event models COG interactions with multiple Golgi SNAREs and states that individual functions remain incompletely characterized. The COG8 physical-entity chain supports the compartment annotation; direct COG8 SNARE binding is separately grounded in PMID:23462996.

Summary: COG8 is represented at the Golgi membrane in intra-Golgi retrograde tethering.

Reason: The event models COG cooperation with CUX1 and GOLGA5 on retrograde intra-Golgi carriers. Its proposed SNARE combinations and RAB cascade remain model-level context, not direct COG8 binding or Rab-regulator assays.

Summary: COG8 is represented at the Golgi membrane in intra-Golgi membrane fusion.

Reason: The event places COG in the context of cis-SNARE formation and membrane fusion, after which tethering factors recycle. It supports the original Golgi membrane compartment but does not attribute fusogenic activity to COG8.

Summary: COG8 is represented at the Golgi membrane in COG-complex Rab associations.

Reason: The event summarizes Rab interactions with COG components and notes that some functional significance is unresolved. Its COG8 entity chain supports membrane localization, without assigning every Rab contact or GTPase regulation to this subunit.

Summary: The inherited intra-Golgi recycling role is retained.

Reason: The human target has a current experimental annotation to this same term and is an integral component of the Golgi recycling machinery. Target self-inclusion is legitimate descendant evidence, not circularity. The historical PAINT tree and IBD placement were not inspected. Endosome-to-TGN experiments are assessed separately and are not substituted for the COPI intra-Golgi definition.

Summary: COG8 participates in COG-dependent Golgi recycling.

Reason: Retain the original experimental curation in the established COG recycling context. The full source directly includes human COG8 knockout, rescue of the glycosylation defect and altered Golgi morphology. It selects representative knockouts for some toxin assays: the inspected text names COG8 for toxin binding but does not identify it in every SubAB transport measurement. Thus the annotation is retained holistically, without claiming that an individually resolved COG8 COPI-flux assay was verified. The distinct endosome-to-TGN assays are not equivalent to this narrower term.

Summary: COG8 binds the RGK-family partner GEM in binary screening.

Reason: Three exact positive two-hybrid records identify canonical human GEM. The official partner belongs to the small-GTPase RGK family, supporting small GTPase binding. Its older function comment reports undetectable intrinsic hydrolysis, so this refinement uses the established family class rather than claiming a measured hydrolysis rate, nucleotide preference, or COG8 GEF/GAP activity. Screen-level orthogonal validation is not asserted for this individual pair.

Proposed replacements: small GTPase binding

Summary: The MYEF2 isoform-2 interaction is retained at its tested-product scope.

Reason: Three positive binary records map to Q9P2K5-2; two specify partner residues 366–576, whereas the validated record gives no fragment boundary. The reconstructed 576-residue isoform retains an RNA-recognition region. The official DNA-repressor description is by similarity and does not resolve that tested fragment's transcription-factor activity. Retain the positive pair without a specific transcriptional binding class or a native COG8 transcription mechanism.

Summary: COG8 binds the mapped 699-residue HGS product.

Reason: The three positive binary records map to the exact A0A0S2Z4Q4 sequence, distinct from canonical 777-residue HGS in the adjacent source assertion. Preserve both source entities. HGS sorting/adaptor biology does not by itself establish a catalytic binding class, assembled-ESCRT binding, or a physiological COG8-HGS trafficking complex.

Summary: COG8 binds canonical HGS in the isoform-interaction study.

Reason: Three positive binary records map to canonical O14964, independently of the shorter HGS product. The source retested isoform pairs and verified clone identities; sampled mammalian complementation is not assigned to this pair. Retain this canonical-product interaction without transferring HGS sorting functions to COG8.

Summary: COG8 binds the DNA-binding transcription factor KLF11.

Reason: Three positive two-hybrid records match canonical human KLF11. Its official experimental functional annotation and curated RNA polymerase II-specific transcription-factor activity establish the partner class, supporting the more specific binding term. The assay does not establish COG8-mediated transcription, a promoter context, or a physiological nuclear interaction.

Proposed replacements: RNA polymerase II-specific DNA-binding transcription factor binding

Summary: COG8 binds the tested tail of KIF1B isoform 2.

Reason: All three positive records specify KIF1B-2 residues 1192–1770. This tail belongs to the verified kinesin product, whose N-terminal motor domain is absent from the tested fragment. Kinesin binding expresses the partner identity and the measured tail contact; it does not claim that the fragment generates force, that full-length motor activity was assayed, or that COG8 drives axonal transport.

Proposed replacements: kinesin binding

Summary: The positive transthyretin pair is retained.

Reason: The three exact records map to the 147-residue TTR precursor without a processed-chain feature. TTR is a soluble carrier rather than a transmembrane transporter; the assay does not establish a mature tetrameric contact, native secretory-pathway encounter, or narrower supported binding class.

Summary: The NDUFS1 isoform-2 pair is retained.

Reason: Three exact positive records identify the 741-residue NDUFS1-2 product. Its official role is electron transfer and assembly within respiratory complex I. This pair does not establish binding to the assembled enzyme or resolve an autonomous catalytic activity of the tested subunit, so the positive interaction is retained without a mitochondrial enzymatic mechanism for COG8.

Summary: COG8 binds several tested huntingtin fragments.

Reason: Twelve positive records preserve four fragment groups: 1221–1939, 1–511, 1–546 and 1932–2642. The N-terminal groups also carry explicit mutation features around residues 18–38, including altered interaction strength. These records are not uniformly wild-type full-length HTT assays. Retain the source-specific positives without inferring a native full-length interaction or a COG8 neurodegeneration mechanism.

Summary: COG8 binds the BAG6 isoform-2 chaperone product.

Reason: The three positive records match the reconstructed 1126-residue BAG6-2 protein, with a six-residue internal splice deletion. Official functional evidence identifies BAG6 as an ATP-independent chaperone that keeps hydrophobic or unfolded clients soluble for transport. The GO definition includes correct folding or transport, supporting this partner-class refinement without assigning chaperone activity to COG8 or claiming an isoform-specific folding assay.

Proposed replacements: protein-folding chaperone binding

Summary: The NUP58 interaction is retained.

Reason: Three exact positive two-hybrid records identify canonical NUP58. Its nuclear-pore membership does not turn an isolated subunit pair into binding to the assembled pore, nor establish a COG8 nuclear-transport activity. The supported association remains non-core.

Summary: Human Protein Atlas microscopy supports Golgi localization.

Reason: The authenticated HPA XML assigns supported Golgi-apparatus immunofluorescence in A-431, MCF-7 and U2OS cells. The source metadata and its current GOA link were checked; image pixels were not independently scored. This is consistent with direct human COG8 trafficking studies.

Summary: COG8 helps maintain Golgi organization.

Reason: The source directly reports dilated Golgi membranes arranged in altered mini-stacks in human COG8-knockout HEK293T cells. This target-specific morphological phenotype complements the demonstrated role of COG8 in COG assembly and SNARE organization; it is not an assertion that every marker assay in the paper was performed in COG8 cells.

Summary: The membrane-proteomics assertion is retained without added compartment specificity.

Reason: The original study analyzed membrane preparations from the NK-like YTS cell line, including peripheral-protein stripping and mass spectrometry. Its exact COG8 target-table entry remains unverified. Independent human COG8 Golgi-associated localization and carrier-recruitment experiments make membrane association biologically well supported, so the original broad HDA assertion is retained non-core with curator deference. This does not verify detection in that particular YTS preparation or imply integral membrane insertion; no negative inference is drawn from the missing table.

Summary: The inherited COG-complex membership is well supported.

Reason: Current records resolve the plant, fly, yeast and human source entities and recover same-term experimental or curated assertions. Independent human assembly experiments directly place COG8 in the complex. The extant list is not a reconstruction of the historical PAINT decision; the original node is retained without a claim that its placement was inspected.

Summary: Human COG8 incorporates into the COG complex.

Reason: Translation of human COG subunits in rabbit reticulocyte lysate and co-immunoprecipitation show COG8 incorporation with COG5, COG6 and COG7. The two-subunit assay did not detect a COG8 binary partner, and omitting COG1 did not prevent this incorporation. This source establishes assembly membership, not the later COG1–COG8 bridge model.

Summary: The InterPro mapping identifies the COG8 complex subunit.

Reason: Both original InterPro identifiers are preserved and agree with the official COG8 family identity and independent human complex-assembly data. Retain the component assertion; the mapping is not used to infer autonomous tethering, fusion or catalysis.

Summary: The conserved Dor1/COG8 component assignment is retained.

Reason: The available original abstract identifies yeast Dor1 in an eight-protein Golgi transport complex and higher-eukaryote homologs. It is not described as a direct human assay. The preserved human NAS membership is independently supported by original human COG8 assembly experiments.

Summary: COG8 is an integral participant in COG assembly.

Reason: The human HEK293T study directly includes COG8 knockout and its effects on partner levels. Independent COG8-depleted HeLa co-immunoprecipitation shows separation of COG subassemblies and altered SNARE-complex assembly. These support complex membership and architecture, without converting loss-of-function phenotypes into a claim of autonomous COG8 fusion.

Summary: COG8 is represented at the TGN membrane in endosomal carrier capture at the TGN.

Reason: The model places COG and GARP with RAB6 and SNAREs during endosomal carrier capture, while explicitly lacking evidence that COG and GARP capture one vesicle together. COG8 membership supports the TGN membrane location, not a demonstrated joint COG–GARP tether or all modeled Rab contacts.

Summary: COG8 is represented at the TGN membrane in endosome-derived vesicle fusion at the TGN.

Reason: The event describes fusion after TGN capture and states that its details remain incompletely established. The COG8-containing entity supports TGN membrane association rather than isolated COG8 fusion activity.

Summary: Human COG8 binds several Golgi SNARE proteins.

Reason: Directed pairwise assays of human COG subunits with Golgi SNAREs identify COG8 interactions with STX5, STX6, STX16 and GS27, corroborated by co-immunoprecipitation. Deletion mapping places the STX16 interaction in its C-terminal SNARE-domain-containing region. This specific molecular activity is independently evidenced and already central to the core function; the original generic-binding assertions concern other partners and remain intact. The user nonredundancy instruction is framed for NEW process terms, rather than excluding a distinct, directly demonstrated MF because an unrelated partner-specific assertion uses its broad parent. The broad SNARE term covers the demonstrated syntaxin and non-syntaxin partners without adding nested binding terms. Binding alone does not establish carrier recruitment, membrane fusion or a physiological interface for every fragment.

Summary: COG8 contributes carrier recognition and capture to endosome-to-TGN retrograde transport.

Reason: Human COG8 retargeted to mitochondria redirects STX16-bearing, internalized Shiga-toxin-containing carriers; binding-retaining COG8 truncations fail to redirect them. The assay retains lobe B context and supports COG8-dependent spatial recognition/capture rather than isolated COG8 tether sufficiency. Independently, COG8 depletion impairs TGN38 return to the TGN in human HeLa cells and alters retrograde cargo distributions in patient fibroblasts. COG8 performs machinery work rather than being the transported substrate. This route is distinct from the retained COPI intra-Golgi process and Golgi organization; typed ancestor/descendant checks against every original, replacement and core term find no overlap. Role-matched GARP components VPS53 and VPS54 carry this exact process in current GO records, with qualifier-specific differences preserved in the notes; no indexed GO-CAM target model gives a competing role. No autonomous fusion or universal product activity is implied.