COA7

Summary: Specify the mitochondrial complex-IV assembly process.

Reason: Human COA7 depletion and rescue support complex-IV assembly, and direct SCO redox chemistry supplies a performed step in that process. The broader inherited function is sound; the mitochondrial replacement increases specificity for this human assertion. TERM_SCOPING_PROBLEM records that target-level scope judgment, not a demonstrated error in historical PAINT node placement. The independent IMP row is retained with its own original source; convergent replacements do not collapse source assertions or create a NEW process.

Proposed replacements: mitochondrial respiratory chain complex IV assembly

Summary: Inherited reductase activity agrees with direct human COA7 experiments.

Reason: Purified reduced heme-COA7 reduces disulfides in soluble SCO1 and SCO2 domains. Human knockout/rescue experiments also shift their cellular redox states. The physiological upstream electron donor and repeated catalytic cycling remain unresolved; the positive reductase capability is nevertheless directly supported.

Summary: The inherited activity location is the mitochondrial intermembrane space.

Reason: Human fractionation, swelling/protease accessibility and super-resolution localization support IMS residence, consistent with the assembly chemistry. The earlier matrix interpretation is separately recorded as disputed rather than silently omitted. The original is_active_in qualifier remains unchanged.

Summary: The UniProt subcellular mapping agrees with direct IMS localization.

Reason: The current mapping uses SL-0169 and the protected UniProt record identifies the mitochondrial intermembrane space. Human biochemical and imaging experiments independently support that compartment.

Summary: Retain the historical unscoped positive binding assertion from HuRI.

Reason: Five current partner-scoped assertions map to positive array, prey-pooling and validated yeast two-hybrid records. Global orthogonal-validation rates are not individual-pair validation. The original aggregate has no partner field and is preserved without narrowing it to one class. Separate current imports permit partner-specific review. These observations do not establish a native IMS mechanism, but that limitation does not negate positive binding. The campaign retains supported generic associations as non-core. Retired marks this unscoped historical aggregate as replaced in the current source projection by the separately preserved partner-scoped assertions; it does not negate the interactions or delete the original record.

Summary: Retain the historical unscoped positive binding assertion from the neurodegeneration interactome.

Reason: Twelve current partner-scoped assertions map to positive array, prey-pooling and validated yeast two-hybrid records. The paper uses full-length and fragment constructs; its DULIP subset is not assigned to every COA7 pair. The original aggregate has no partner field and is preserved without narrowing it to one class. Separate current imports permit partner-specific review. These observations do not establish a native IMS mechanism, but that limitation does not negate positive binding. The campaign retains supported generic associations as non-core. Retired marks this unscoped historical aggregate as replaced in the current source projection by the separately preserved partner-scoped assertions; it does not negate the interactions or delete the original record.

Summary: Retain the historical unscoped positive binding assertion from BioPlex.

Reason: Three exact anti-tag coimmunoprecipitation records identify human HCT116 associations with ACOT8, DPP8 and BKGD/C11orf54. The original aggregate has no partner field and is preserved without narrowing it to one class. Separate current imports permit partner-specific review. These observations do not establish a native IMS mechanism, but that limitation does not negate positive binding. The campaign retains supported generic associations as non-core. Retired marks this unscoped historical aggregate as replaced in the current source projection by the separately preserved partner-scoped assertions; it does not negate the interactions or delete the original record.

Summary: Retain the historical unscoped positive binding assertion from the U2OS AP-MS network.

Reason: The original linked NDEx network directly contains COA7 with ACOT8, DPP8 and C11orf54. Each partner is marked as bait and COA7 as prey; these are AP-MS edges, not simply image-derived proximity predictions. The original aggregate has no partner field and is preserved without narrowing it to one class. Separate current imports permit partner-specific review. These observations do not establish a native IMS mechanism, but that limitation does not negate positive binding. The campaign retains supported generic associations as non-core. Retired marks this unscoped historical aggregate as replaced in the current source projection by the separately preserved partner-scoped assertions; it does not negate the interactions or delete the original record.

Summary: Human perturbation supports mitochondrial complex-IV assembly.

Reason: The original complete abstract reports human C1orf163/RESA1 depletion with strongest assembly defects in complex IV; its full paper remains unrecovered. Later human knockout/rescue and SCO redox experiments independently identify COA7 participation in assembly beyond necessity alone. Complex-I changes are preserved as observations without adding a separate process.

Summary: Reduced heme-bound COA7 directly reduces SCO protein disulfides.

Reason: Recombinant human COA7 expressed in E. coli was GST-cleaved, heme-loaded and reduced; excess dithionite was removed before reaction with soluble SCO1 128–301 or SCO2 96–266. Mass shifts identify disulfide reduction. SCO1 controls include hemin alone, apo protein, heme-ligand mutant and cytochrome c; SCO2 SI S18 separately compares WT and Y137C. Cellular rescue supports the redox role, while electron-donor identity and turnover kinetics remain open.

Summary: Retain the original experimentally supported IMS activity compartment.

Reason: The authentic 2013 abstract explicitly identifies a soluble IMS protein in human mitochondria. Although the original full text was not recovered, later human swelling/protease and imaging experiments independently establish the same localization. Preserve the source is_active_in qualifier.

Summary: The mitochondrial proteome contains the exact COA7 target.

Reason: Original MitoCoP Table S1 lists COA7/Q96BR5 in the all-protein table (row 222) and high-confidence mitochondrial set (row 197). Human subcellular proteomics supports the broad mitochondrial annotation. This source does not independently resolve IMS versus matrix; the more precise IMS assignment rests on other experiments.

Summary: Specify binding to the CHCHD4/MIA40 oxidoreductase.

Reason: The human-cell primary demonstrates a cysteine-dependent COA7–CHCHD4 complex and COA7 import as a CHCHD4 substrate. The partner is an enzyme, making enzyme binding informative without assigning CHCHD4 import or oxidative-folding catalysis to COA7.

Proposed replacements: enzyme binding

Summary: Direct human experiments localize soluble COA7 to the IMS.

Reason: HEK293 mitochondrial swelling releases COA7 and exposes it to protease while matrix controls remain protected; endogenous and tagged HeLa imaging independently favors the IMS marker. This later study explicitly addresses the earlier matrix interpretation.

Summary: Retain the positive non-core association with SNRPB isoform SM-B.

Reason: Three exact positive yeast two-hybrid records (array, prey pooling and validated retest) support this pair. The exact P14678-2 product reconstructs to 231 residues and retains the Sm domain. This positive association with a spliceosomal protein does not resolve a more specific binding MF or assign RNA splicing to COA7.

Summary: Specify the partner class for SUOX.

Reason: Three exact positive yeast two-hybrid records (array, prey pooling and validated retest) support this pair. P51687 is the sulfite oxidase enzyme. The positive pair supports binding to an enzyme without showing that COA7 regulates its chemistry or serves as its physiological electron donor. The official SUOX record carries experimentally supported catalysis and an IMS localization by similarity; the partner is not a generic nonmitochondrial noise example.

Proposed replacements: enzyme binding

Summary: Specify the partner class for TLE5 isoform 2.

Reason: Three exact positive HuRI yeast two-hybrid records support the COA7–Q08117-2 pair. The official definition of GO:0001222 describes binding to a transcription corepressor; it does not require the binding protein itself to participate in transcriptional repression. Experimentally supported official TLE5 function and its current experimental corepressor annotation establish the partner class. The exact 264-residue isoform 2 replaces residues 1–9 of the displayed sequence with its alternative N terminus and retains the shared remainder. This supports a qualified partner-class inference, not an independently measured isoform-specific repression activity, a mapped binding interface, or a native transcriptional role for COA7. Partner-function primary papers were not newly read for this reassessment.

Proposed replacements: transcription corepressor binding

Summary: Specify binding to the ENKD1 cytoskeletal regulatory protein.

Reason: Three exact positive HuRI yeast two-hybrid records support binding to the mapped 346-residue ENKD1 product. The experimentally supported official ENKD1 function describes regulation of microtubule organization and stability, including stabilization of astral microtubules and spindle orientation. These positive regulatory capabilities support GO:0005519 cytoskeletal regulatory protein binding, a more specific child of GO:0008092, rather than relying on component localization alone. The definition identifies the bound partner class; it does not require COA7 to regulate or join the cytoskeleton. No native COA7 cytoskeletal role, tubulin interaction or contacted domain is inferred. The partner-function papers cited by the official record were not newly read; the class inference rests on the experimentally supported official record and the exact positive product mapping.

Proposed replacements: cytoskeletal regulatory protein binding

Summary: Retain the positive non-core association with GPSM3.

Reason: Three exact positive yeast two-hybrid records (array, prey pooling and validated retest) support this pair. Q9Y4H4 modulates heterotrimeric G-protein subunits; it is not itself a small GTPase. The positive pair is retained without substituting a GTPase-binding term merely from the partner’s regulatory role.

Summary: Retain the positive non-core association with PMP22.

Reason: Three exact positive yeast two-hybrid records (array, prey pooling and validated retest) support this pair. The literal partner is A0A6Q8PF08. Its membrane-protein identity does not establish an enzymatic, receptor or transporter class that would justify a more specific binding claim here.

Summary: Retain the positive non-core association with DMWD fragment.

Reason: Three exact positive yeast two-hybrid records (array, prey pooling and validated retest) support this pair. The exact G5E9A7 record specifies residues 316–649. This positive fragment interaction does not identify a supported specific functional binding class or demonstrate a full-length physiological complex.

Summary: Retain the positive non-core association with PRPF40A isoform 2 fragment.

Reason: Three exact positive yeast two-hybrid records (array, prey pooling and validated retest) support this pair. O75400-2 reconstructs to 930 residues, and the assay records residues 716–930. RNA-processing and cytoskeletal-regulatory descriptions of the full protein do not establish a cytoskeletal-component class for this tested fragment.

Summary: Specify binding to a cytoskeletal regulatory protein.

Reason: Three exact positive neurodegeneration-screen yeast two-hybrid records establish the gelsolin association. All recorded sequence fingerprints match the 782-residue displayed secreted/plasma precursor, not the distinct 731-residue cytoplasmic isoform 2. The official gelsolin function includes actin capping, nucleation and severing: these positive actin-remodeling capabilities support the regulatory-protein class required by GO:0005519. This more specific term is a child of GO:0008092; cytoplasmic isoform-specific localization is not used to deny the shared regulatory capability of the mapped precursor. The refinement identifies the bound protein class, not COA7-driven cytoskeletal reorganization, mature plasma processing in yeast, native colocalization or a particular binding interface. Partner-function primary papers were not newly read; the functional class rests on the experimentally supported official record.

Proposed replacements: cytoskeletal regulatory protein binding

Summary: Retain the positive non-core association with LAMP2B.

Reason: Three exact positive yeast two-hybrid records (array, prey pooling and validated retest) support this pair. The exact P13473-2 product reconstructs to 410 residues and corresponds to LAMP2B. Binding does not demonstrate a particular lysosomal receptor interface or transfer the partner’s autophagy role to COA7.

Summary: Specify the partner class for FGFR3.

Reason: Three exact positive neurodegeneration-screen yeast two-hybrid records support the mapped P22607 receptor pair. The current official GO:0005104 definition is binding to an FGFR; its typed ancestry places it under receptor/protein binding and imposes no separate requirement for ligand agonism or FGF-pathway membership. Official FGFR3 identity establishes the receptor class. Retain the assay-scoped refinement without inferring receptor activation, plasma-membrane localization of COA7, a mature processed receptor state in yeast, or the contacted domain. Native physiological relevance remains unresolved.

Proposed replacements: fibroblast growth factor receptor binding

Summary: Specify the partner class for HMOX2.

Reason: Three exact positive yeast two-hybrid records (array, prey pooling and validated retest) support this pair. P30519 is a heme oxygenase enzyme. The positive pair supports enzyme binding without showing COA7-driven heme degradation, heme transfer or regulation of HMOX2. The official HMOX2 record assigns experimentally supported catalytic activity to the protein and its soluble form. No catalytic activity is assigned to COA7 by this refinement.

Proposed replacements: enzyme binding

Summary: Specify the partner class for CASP6.

Reason: Three exact positive yeast two-hybrid records (array, prey pooling and validated retest) support this pair. P55212 is a caspase protease precursor. Binding to that authenticated protein class does not require or demonstrate activation of its catalytic sites, cleavage of COA7 or an apoptosis role for COA7. This is binding to the authenticated caspase-6 protein class; the screen does not establish maturation, active-site occupancy or proteolytic activity of the bound precursor.

Proposed replacements: protease binding

Summary: Specify the partner class for RAN.

Reason: Three exact positive yeast two-hybrid records (array, prey pooling and validated retest) support this pair. P62826 is a small monomeric GTPase. The positive pair supports this binding class without demonstrating GEF, GAP or nuclear-transport activity by COA7. The official RAN record establishes nucleotide switching and GTP hydrolysis. GO:0031267 describes binding to that small monomeric GTPase, without requiring a GEF/GAP function of the binding partner.

Proposed replacements: small GTPase binding

Summary: Retain the positive non-core association with CYCS.

Reason: Three exact positive yeast two-hybrid records (array, prey pooling and validated retest) support this pair. P99999 is an electron carrier. Association does not establish transfer of electrons between this pair or justify an enzyme-binding refinement simply because the partner participates in respiratory chemistry.

Summary: Retain the positive non-core association with SPRED1.

Reason: Three exact positive yeast two-hybrid records (array, prey pooling and validated retest) support this pair. Q7Z699 is a signaling regulator. Binding alone does not establish that this partner is itself a GAP or kinase, and no more specific tested interface is identified.

Summary: Retain the positive non-core association with GW128.

Reason: Three exact positive yeast two-hybrid records (array, prey pooling and validated retest) support this pair. The literal partner Q9Y649 is an unreviewed GW128 entry with no sufficiently resolved functional class in the authenticated record. The positive interaction is retained without inventing an enzyme or cytoskeletal identity.

Summary: Specify the partner class for ACOT8.

Reason: An exact HCT116 anti-tag coimmunoprecipitation record supports this pair. O14734 is an acyl-CoA thioesterase; the authenticated class supports enzyme binding, without demonstrating catalytic regulation or a native mitochondrial assembly role for this pair. The retained term is limited to the experimentally observed association with an independently supported catalytic protein class; it does not resolve whether contact is direct within the recovered complex.

Proposed replacements: enzyme binding

Summary: Specify the partner class for DPP8.

Reason: An exact HCT116 anti-tag coimmunoprecipitation record supports this pair. Q6V1X1 is a dipeptidyl peptidase. Association supports protease binding, not evidence that COA7 is cleaved or that it regulates inflammasome activity. The retained term is limited to the experimentally observed association with an independently supported catalytic protein class; it does not resolve whether contact is direct within the recovered complex.

Proposed replacements: protease binding

Summary: Specify the partner class for BKGD/C11orf54.

Reason: An exact HCT116 anti-tag coimmunoprecipitation record supports this pair. Q9H0W9 is now named beta-keto L-gulonate decarboxylase; the network’s C11orf54 label is an authenticated alias. Its enzyme class supports this refinement without importing its metabolic reaction into COA7 biology. The retained term is limited to the experimentally observed association with an independently supported catalytic protein class; it does not resolve whether contact is direct within the recovered complex.

Proposed replacements: enzyme binding

Summary: Specify the partner class for ACOT8.

Reason: Original U2OS AP-MS network data directly corroborate the current GOA pair. The U2OS network contains the exact ACOT8–COA7 edge, with ACOT8 marked as bait and COA7 as prey. Official O14734 enzyme identity supports the class refinement; coisolation is not proof of a binary interface or catalytic regulation. The retained term is limited to the experimentally observed association with an independently supported catalytic protein class; it does not resolve whether contact is direct within the recovered complex.

Proposed replacements: enzyme binding

Summary: Specify the partner class for DPP8.

Reason: Original U2OS AP-MS network data directly corroborate the current GOA pair. The U2OS network contains the exact DPP8–COA7 edge, with DPP8 marked as bait and COA7 as prey. Official Q6V1X1 peptidase identity supports protease binding at association scope, without demonstrating COA7 cleavage. The retained term is limited to the experimentally observed association with an independently supported catalytic protein class; it does not resolve whether contact is direct within the recovered complex.

Proposed replacements: protease binding

Summary: Specify the partner class for BKGD/C11orf54.

Reason: Original U2OS AP-MS network data directly corroborate the current GOA pair. The U2OS network’s C11orf54–COA7 edge maps to Q9H0W9, now named BKGD; C11orf54 is bait and COA7 prey. The current enzyme class supports binding specificity without assigning this enzyme’s reaction to COA7. The retained term is limited to the experimentally observed association with an independently supported catalytic protein class; it does not resolve whether contact is direct within the recovered complex.

Proposed replacements: enzyme binding

Summary: Add directly measured heme binding as a cofactor capability.

Reason: Purified human COA7 binds hemin with a micromolar fitted affinity and approximately one heme per protein; axial-ligand substitutions alter the interaction. This supplies nonredundant MF coverage within the reductase core, not a separate heme-delivery function. Typed ancestry and descendants exclude overlap with all retained source terms and final replacements; no NEW process is proposed.