EIF3F

Summary: The rule supports initiation more broadly than the large-subunit joining step.

Reason: UR000106280 genuinely applies to the human HAMAP MF_03005 record and emits this term. However, GO:0001732 specifically describes large-subunit joining and initiation-factor release. The experimentally supported EIF3F contribution is assembly and function of eIF3-containing preinitiation complexes; the rule does not independently establish an EIF3F-specific role at that later joining step. Use cytoplasmic translational initiation without denying the literal rule output.

Proposed replacements: cytoplasmic translational initiation

Summary: Refine the review-level initiation claim to the experimentally established cytoplasmic process.

Reason: The cited synthesis describes initiation-factor organization. Its available abstract does not resolve the particular large-subunit joining/release step required by GO:0001732. Independent human eIF3 assembly and 48S reconstitution support participation in cytoplasmic translational initiation, a broader process that does not claim EIF3F catalyses subunit joining. This is a biological generalization, not a wrong-reference allegation.

Proposed replacements: cytoplasmic translational initiation

Summary: EIF3F contributes structurally to cytoplasmic translation initiation.

Reason: The matching UniRule predicate identifies the conserved eIF3F protein. Human recombinant eIF3 complexes containing F support 48S formation, and native human eIF3 contains a defined F:H:M module. This establishes participation through assembly of the initiation machinery rather than merely necessity after depletion.

Summary: Retain the phylogenetically grounded contribution to initiation-factor activity.

Reason: The human leaf descends from the literal PTN000056988 IBD for GO:0003743 with contributes_to. Human and mouse experimental annotations ground the assertion; the target appearing among donors is legitimate. Human reconstitution supports activity of assembled eIF3 containing F, not an isolated F initiation assay.

Summary: EIF3F contributes to the activity of native eIF3.

Reason: The cited abstract identifies all 13 subunits of native HeLa eIF3 as stoichiometric components. Complete operative text was not recovered from the publisher, so no source-specific F-minus activity experiment is inferred. Independent human reconstitution and native-complex architecture support this explicitly contributes_to molecular function.

Summary: The native F-containing eIF3 assembly supports initiation-factor activity.

Reason: The source purified human HeLa cytosolic eIF3 and resolved an F:H:M module and F:H/F:M subcomplexes. This structural participation supports the imported contributes_to relation; mass spectrometry alone is not an assay of autonomous EIF3F initiation activity. Functional reconstitution provides independent complex-level support.

Summary: Human recombinant eIF3 containing F supports initiation.

Reason: Human 11- and six-subunit eIF3 preparations support 48S toeprints. The native comparator is mouse Krebs-2 eIF3; the five-subunit control lacks C, whereas F-minus material had insufficient purified yield. Preserve contributes_to without claiming a matched purified F-minus activity loss or an isolated F assay.

Summary: EIF3F participates in the RNA-regulating initiation-factor complex.

Reason: Human eIF3 was isolated from 293T cells and used in RNA-associated and initiation experiments. Denaturing PAR-CLIP assigns direct crosslinks to A/B/D/G, not F; the F-containing assembly nevertheless provides the imported contribution to initiation-factor activity. No direct RNA-binding activity of isolated EIF3F is inferred.

Summary: Retain the conserved initiation-factor function with explicit complex context.

Reason: The aggregate IEA channels resolve to mouse Q9DCH4/ENSMUSP00000033342, the eIF3F InterPro family and an applicable UniRule. Human functional reconstitution independently supports the function. The imported enables qualifier is preserved, but this review describes activity in assembled eIF3 and does not interpret that qualifier as evidence for an isolated F assay.

Summary: Human p47/EIF3F is an initiation-factor subunit.

Reason: The source abstract identifies human p47 cloning and association with eIF3 through anti-p170 immunoprecipitation; its p66 RNA-binding result is not transferred to p47. Later human reconstitution supports initiation-factor function in the assembled complex. Retain the imported enables qualifier without inventing an autonomous F activity measurement.

Summary: Retain deubiquitinase activity without assigning thiol-dependent chemistry.

Reason: The source genuinely combines mouse Q9DCH4 transfer with EC:3.4.19.12, whose current ec2go mapping specifies cysteine-type deubiquitinase activity. PMID:21124883 contains positive human Ub-GFP cleavage and a partial mouse Ub-vinyl-sulfone adduct, with mouse mutant/rescue experiments distinct from the human assay. These findings support DUB activity, but the inspected experiments do not establish the human catalytic nucleophile. Generalize to GO:0101005 rather than reject activity or infer chemistry from the MPN fold.

Proposed replacements: deubiquitinase activity

Summary: Refine the established eIF3 assembly interaction to initiation-factor binding.

Reason: The literal record is a human EIF3F–EIF3M pooled two-hybrid pair. Independent native human F:M and F:H:M subcomplexes establish an initiation-complex role for this contact. GO:0031369 names binding to a polypeptide factor involved in translation initiation; it does not transfer the partner’s catalytic activity.

Proposed replacements: translation initiation factor binding

Summary: Retain the coronavirus spike association as a non-core interaction.

Reason: The literal partner is IBV spike P11223. The original reports human EIF3F association with an IBV spike 19–400 fragment and full-length co-expression constructs, with infected Vero-cell co-immunoprecipitation providing a separate context. Preserve the positive interaction without inferring that EIF3F is a viral receptor or that each construct establishes the same native interface.

Summary: Retain the experimentally supported SARS spike interaction.

Reason: The literal partner is SARS-CoV spike P59594, including the 20–404 construct used in the screen. The source follows up with co-immunoprecipitation and purified GST-fusion sedimentation; these support an interaction in the tested constructs. No specific native receptor, catalytic or regulatory function of EIF3F follows from this pair.

Summary: EIF3F binds the FBXO32 ubiquitin-ligase machinery as its substrate.

Reason: Human skeletal-muscle library screening and reciprocal in-vitro binding map EIF3F interaction with FBXO32/MAFbx, while recombinant SCF(FBXO32) ubiquitinates EIF3F. Mouse C2C12 atrophy experiments are a separate cellular context. FBXO32 is the substrate-recognition component of this E3 machinery; the refinement describes EIF3F binding that machinery, not EIF3F catalysing ubiquitination. Exact historical inserts are not equated with all natural products.

Proposed replacements: ubiquitin protein ligase binding

Summary: Refine the established eIF3 assembly interaction to initiation-factor binding.

Reason: Native human HeLa eIF3 mass spectrometry resolves an F:H binary subcomplex as well as F:H:M; this is a defined assembly interaction with EIF3H. GO:0031369 names binding to a polypeptide factor involved in translation initiation; it does not transfer the partner’s catalytic activity.

Proposed replacements: translation initiation factor binding

Summary: Refine the established eIF3 assembly interaction to initiation-factor binding.

Reason: The source describes EIF3B association with F in immunoprecipitation during its dissection of native human eIF3. The separately unread supplementary figure is not treated as a purified binary assay. GO:0031369 names binding to a polypeptide factor involved in translation initiation; it does not transfer the partner’s catalytic activity.

Proposed replacements: translation initiation factor binding

Summary: Refine the established eIF3 assembly interaction to initiation-factor binding.

Reason: Native human HeLa eIF3 mass spectrometry resolves F:M and F:H:M subcomplexes, establishing the EIF3M assembly interaction. GO:0031369 names binding to a polypeptide factor involved in translation initiation; it does not transfer the partner’s catalytic activity.

Proposed replacements: translation initiation factor binding

Summary: Refine the established eIF3 assembly interaction to initiation-factor binding.

Reason: RNase-treated human HeLa cytoplasmic co-immunoprecipitation and mass spectrometry recover eIF3 subunits, with reciprocal EIF3B immunoprecipitation. This supports association within established eIF3, not direct contact between every recovered protein. GO:0031369 names binding to a polypeptide factor involved in translation initiation; it does not transfer the partner’s catalytic activity.

Proposed replacements: translation initiation factor binding

Summary: Retain the human GLE1 association without a new functional subtype.

Reason: The source reports human GLE1-GFP and EIF3F-HA co-immunoprecipitation in HeLa cells. Its yeast translation-termination and Dbp5 mechanism is not transferred to human EIF3F, which is absent from budding yeast. The association is supported, but it does not independently establish a specific EIF3F regulatory activity.

Summary: EIF3F interacts with the processed CDK11 kinase as a substrate.

Reason: In human A375 cells, apoptotic caspase-processed CDK11p46 is immunopurified and phosphorylates recombinant EIF3F constructs; Thr119 and prior Ser46 phosphorylation context is tested separately from eIF3 assembly. The source partner P21127/CDK11B is preserved, while the assayed processed kinase is not equated with an intact current isoform or every insert. GO:0019901 describes protein-kinase binding and does not imply that EIF3F is itself a kinase.

Proposed replacements: protein kinase binding

Summary: Refine the established eIF3 assembly interaction to initiation-factor binding.

Reason: The human A375 experiments recover EIF3F with EIF3B and test phosphorylation-dependent changes in this association. This is a contact within the established initiation assembly, distinct from the CDK11 substrate interaction. GO:0031369 names binding to a polypeptide factor involved in translation initiation; it does not transfer the partner’s catalytic activity.

Proposed replacements: translation initiation factor binding

Summary: Refine the established eIF3 assembly interaction to initiation-factor binding.

Reason: The accessible source reports ion-mobility analysis of human eIF3 subcomplex organization. Its detailed target contact was not independently recovered, but the earlier directly inspected native-human F:H subcomplex supports the same established assembly interaction. GO:0031369 names binding to a polypeptide factor involved in translation initiation; it does not transfer the partner’s catalytic activity.

Proposed replacements: translation initiation factor binding

Summary: Refine the established eIF3 assembly interaction to initiation-factor binding.

Reason: The accessible ion-mobility source supports human eIF3 architecture. The independently inspected native-human F:M and F:H:M subcomplexes support the imported EIF3M association without inventing an uninspected measurement. GO:0031369 names binding to a polypeptide factor involved in translation initiation; it does not transfer the partner’s catalytic activity.

Proposed replacements: translation initiation factor binding

Summary: Retain the HIV protease association as non-core binding.

Reason: The literal interaction records identify Q9Q2G4 protease constructs, including catalytic-site mutation and tagging features, in co-immunoprecipitation/pull-down assays. The accessible paper’s EIF3D cleavage result is not transferred to EIF3F. Positive association is preserved without an EIF3F protease or viral-entry function.

Summary: Retain the ATXN1 screen interaction with its construct limits.

Reason: The source uses ATXN1 Q32/Q79 constructs and pooled two-hybrid screening; the literal EIF3F pair maps to its interaction table. The exact polyglutamine state and any orthogonal validation subset for this pair are not resolved by the retained record. The positive pair supports non-core binding, not an assigned ATXN1-mediated physiological role.

Summary: Refine the established eIF3 assembly interaction to initiation-factor binding.

Reason: The recombinant human PCI/MPN octamer contains F and H; the structural model places them in the compact eIF3 assembly. The 15-angstrom reconstruction and flexible F tag do not provide an atomically resolved interface, but native human F:H subcomplex evidence independently establishes the role. GO:0031369 names binding to a polypeptide factor involved in translation initiation; it does not transfer the partner’s catalytic activity.

Proposed replacements: translation initiation factor binding

Summary: Refine the established eIF3 assembly interaction to initiation-factor binding.

Reason: Human recombinant core assembly contains F and M, with partial A and C constructs explicitly distinguished. Native human F:M subcomplex evidence independently establishes the role; model proximity is not represented as an atomic interface. GO:0031369 names binding to a polypeptide factor involved in translation initiation; it does not transfer the partner’s catalytic activity.

Proposed replacements: translation initiation factor binding

Summary: Retain the experimentally demonstrated MSH4 interaction.

Reason: Human ovary library screening, reciprocal 293T co-immunoprecipitation and recombinant-protein assays support MSH4–EIF3F binding, mapped to the EIF3F N terminus. The paper’s 106–361 construct notation differs from the current 357-residue displayed product, so insert equivalence is not assumed. Stabilization observations do not establish a new autonomous EIF3F DNA-repair activity or a more specific binding mechanism.

Summary: Retain the GLE1 two-hybrid interaction.

Reason: The source tests human GLE1B and the FinMajor PFQ insertion in two-hybrid assays with EIF3F. The interaction is positive, but the paper’s nuclear-pore and mRNA-export experiments do not by themselves establish a specific export or catalytic function for EIF3F. Preserve the assayed partner variant scope.

Summary: Refine the established eIF3 assembly interaction to initiation-factor binding.

Reason: The source co-immunoprecipitates tagged human EIF3F with endogenous eIF3 components in 293T cells. Its rabbit-reticulocyte S6K assay has an ambiguous 44–47-kDa band and is not used to claim a specific F phosphorylation reaction. GO:0031369 names binding to a polypeptide factor involved in translation initiation; it does not transfer the partner’s catalytic activity.

Proposed replacements: translation initiation factor binding

Summary: Refine the established eIF3 assembly interaction to initiation-factor binding.

Reason: The three literal human EIF3F–EIF3M records use co-immunoprecipitation, protein complementation and MAPPIT. Independent native F:M complex evidence gives this pair an established initiation-assembly role; no unrecorded target-specific assay or insert equivalence is inferred. GO:0031369 names binding to a polypeptide factor involved in translation initiation; it does not transfer the partner’s catalytic activity.

Proposed replacements: translation initiation factor binding

Summary: Refine the established eIF3 assembly interaction to initiation-factor binding.

Reason: BioPlex 2.0 reports affinity-purification mass-spectrometry association of human EIF3F and EIF3H. A native human F:H subcomplex independently supplies the assembly mechanism; the AP-MS edge itself is not called direct binding. GO:0031369 names binding to a polypeptide factor involved in translation initiation; it does not transfer the partner’s catalytic activity.

Proposed replacements: translation initiation factor binding

Summary: Refine the established eIF3 assembly interaction to initiation-factor binding.

Reason: BioPlex 2.0 supplies a positive EIF3F–EIF3M AP-MS association. Native human F:M subcomplexes independently establish the initiation-factor assembly role; abundance or compartment alone is not used to reject the screen. GO:0031369 names binding to a polypeptide factor involved in translation initiation; it does not transfer the partner’s catalytic activity.

Proposed replacements: translation initiation factor binding

Summary: Retain the EML2 screen interaction as non-core binding.

Reason: The literal human EIF3F–EML2 (A0A0C4DGQ7) pair is supported by the HuRI two-hybrid screen and its retesting procedure. The retained record contains construct features: gal4 activation domain:n-n; gal4 dna binding domain:n-n. A current reference-product match does not prove fragment or isoform functional equivalence. Positive assay-context binding is retained; partner identity alone does not establish a specific EIF3F molecular function or native regulatory role, and the pair is not rejected from compartment expectations.

Summary: Retain the CCDC196 screen interaction as non-core binding.

Reason: The literal human EIF3F–CCDC196 (A0A1B0GWI1) pair is supported by the HuRI two-hybrid screen and its retesting procedure. The retained record contains construct features: gal4 activation domain:c-c; gal4 activation domain:n-n; gal4 dna binding domain:n-n. A current reference-product match does not prove fragment or isoform functional equivalence. Positive assay-context binding is retained; partner identity alone does not establish a specific EIF3F molecular function or native regulatory role, and the pair is not rejected from compartment expectations.

Summary: Retain the HGS screen interaction as non-core binding.

Reason: The literal human EIF3F–HGS (O14964) pair is supported by the HuRI two-hybrid screen and its retesting procedure. The retained record contains construct features: gal4 activation domain:n-n; gal4 dna binding domain:n-n. A current reference-product match does not prove fragment or isoform functional equivalence. Positive assay-context binding is retained; partner identity alone does not establish a specific EIF3F molecular function or native regulatory role, and the pair is not rejected from compartment expectations.

Summary: Retain the AGR2 screen interaction as non-core binding.

Reason: The literal human EIF3F–AGR2 (O95994) pair is supported by the HuRI two-hybrid screen and its retesting procedure. The retained record contains construct features: gal4 activation domain:n-n; gal4 dna binding domain:n-n. A current reference-product match does not prove fragment or isoform functional equivalence. Positive assay-context binding is retained; partner identity alone does not establish a specific EIF3F molecular function or native regulatory role, and the pair is not rejected from compartment expectations.

Summary: Retain the COL1A2 screen interaction as non-core binding.

Reason: The literal human EIF3F–COL1A2 (P08123) pair is supported by the HuRI two-hybrid screen and its retesting procedure. The retained record contains construct features: gal4 activation domain:c-c; gal4 dna binding domain:n-n. A current reference-product match does not prove fragment or isoform functional equivalence. Positive assay-context binding is retained; partner identity alone does not establish a specific EIF3F molecular function or native regulatory role, and the pair is not rejected from compartment expectations.

Summary: Retain the RABIF screen interaction as non-core binding.

Reason: The literal human EIF3F–RABIF (P47224) pair is supported by the HuRI two-hybrid screen and its retesting procedure. The retained record contains construct features: gal4 activation domain:c-c; gal4 activation domain:n-n; gal4 dna binding domain:n-n. A current reference-product match does not prove fragment or isoform functional equivalence. Positive assay-context binding is retained; partner identity alone does not establish a specific EIF3F molecular function or native regulatory role, and the pair is not rejected from compartment expectations.

Summary: Retain the SUOX screen interaction as non-core binding.

Reason: The literal human EIF3F–SUOX (P51687) pair is supported by the HuRI two-hybrid screen and its retesting procedure. The retained record contains construct features: gal4 activation domain:c-c; gal4 dna binding domain:n-n. A current reference-product match does not prove fragment or isoform functional equivalence. Positive assay-context binding is retained; partner identity alone does not establish a specific EIF3F molecular function or native regulatory role, and the pair is not rejected from compartment expectations.

Summary: Retain the POU6F2 screen interaction as non-core binding.

Reason: The literal human EIF3F–POU6F2 (P78424) pair is supported by the HuRI two-hybrid screen and its retesting procedure. The retained record contains construct features: gal4 activation domain:n-n; gal4 activation domain:n-n|sufficient binding region:9-691; gal4 dna binding domain:n-n; gal4 dna binding domain:n-n|sufficient binding region:9-691; sufficient binding region:9-691. A current reference-product match does not prove fragment or isoform functional equivalence. Positive assay-context binding is retained; partner identity alone does not establish a specific EIF3F molecular function or native regulatory role, and the pair is not rejected from compartment expectations.

Summary: Retain the RIN1 screen interaction as non-core binding.

Reason: The literal human EIF3F–RIN1 (Q13671) pair is supported by the HuRI two-hybrid screen and its retesting procedure. The retained record contains construct features: gal4 activation domain:n-n; gal4 dna binding domain:n-n. A current reference-product match does not prove fragment or isoform functional equivalence. Positive assay-context binding is retained; partner identity alone does not establish a specific EIF3F molecular function or native regulatory role, and the pair is not rejected from compartment expectations.

Summary: Retain the VEZF1 screen interaction as non-core binding.

Reason: The literal human EIF3F–VEZF1 (Q14119) pair is supported by the HuRI two-hybrid screen and its retesting procedure. The retained record contains construct features: gal4 activation domain:n-n; gal4 dna binding domain:n-n. A current reference-product match does not prove fragment or isoform functional equivalence. Positive assay-context binding is retained; partner identity alone does not establish a specific EIF3F molecular function or native regulatory role, and the pair is not rejected from compartment expectations.

Summary: Retain the RTP5 screen interaction as non-core binding.

Reason: The literal human EIF3F–RTP5 (Q14D33) pair is supported by the HuRI two-hybrid screen and its retesting procedure. The retained record contains construct features: gal4 activation domain:n-n; gal4 dna binding domain:n-n. A current reference-product match does not prove fragment or isoform functional equivalence. Positive assay-context binding is retained; partner identity alone does not establish a specific EIF3F molecular function or native regulatory role, and the pair is not rejected from compartment expectations.

Summary: Retain the PKN1 screen interaction as non-core binding.

Reason: The literal human EIF3F–PKN1 (Q16512) pair is supported by the HuRI two-hybrid screen and its retesting procedure. The retained record contains construct features: gal4 activation domain:n-n; gal4 dna binding domain:n-n. A current reference-product match does not prove fragment or isoform functional equivalence. Positive assay-context binding is retained; partner identity alone does not establish a specific EIF3F molecular function or native regulatory role, and the pair is not rejected from compartment expectations.

Summary: Refine the established eIF3 assembly interaction to initiation-factor binding.

Reason: The HuRI human EIF3F–EIF3M two-hybrid pair is reinforced by independently characterized native F:M subcomplexes. This known assembly role, rather than the partner name alone, supports a specific initiation-factor-binding term. GO:0031369 names binding to a polypeptide factor involved in translation initiation; it does not transfer the partner’s catalytic activity.

Proposed replacements: translation initiation factor binding

Summary: Retain the NUP54 screen interaction as non-core binding.

Reason: The literal human EIF3F–NUP54 (Q7Z3B4) pair is supported by the HuRI two-hybrid screen and its retesting procedure. The retained record contains construct features: gal4 activation domain:n-n; gal4 dna binding domain:n-n|sufficient binding region:1-493; sufficient binding region:1-493. A current reference-product match does not prove fragment or isoform functional equivalence. Positive assay-context binding is retained; partner identity alone does not establish a specific EIF3F molecular function or native regulatory role, and the pair is not rejected from compartment expectations.

Summary: Retain the POGZ screen interaction as non-core binding.

Reason: The literal human EIF3F–POGZ (Q7Z3K3) pair is supported by the HuRI two-hybrid screen and its retesting procedure. The retained record contains construct features: gal4 activation domain:n-n; gal4 dna binding domain:n-n. A current reference-product match does not prove fragment or isoform functional equivalence. Positive assay-context binding is retained; partner identity alone does not establish a specific EIF3F molecular function or native regulatory role, and the pair is not rejected from compartment expectations.

Summary: Retain the C1orf216 screen interaction as non-core binding.

Reason: The literal human EIF3F–C1orf216 (Q8TAB5) pair is supported by the HuRI two-hybrid screen and its retesting procedure. The retained record contains construct features: gal4 activation domain:n-n; gal4 dna binding domain:n-n. A current reference-product match does not prove fragment or isoform functional equivalence. Positive assay-context binding is retained; partner identity alone does not establish a specific EIF3F molecular function or native regulatory role, and the pair is not rejected from compartment expectations.

Summary: Retain the FAM222B screen interaction as non-core binding.

Reason: The literal human EIF3F–FAM222B (Q8WU58) pair is supported by the HuRI two-hybrid screen and its retesting procedure. The retained record contains construct features: gal4 activation domain:n-n; gal4 dna binding domain:n-n. A current reference-product match does not prove fragment or isoform functional equivalence. Positive assay-context binding is retained; partner identity alone does not establish a specific EIF3F molecular function or native regulatory role, and the pair is not rejected from compartment expectations.

Summary: Retain the MYDGF screen interaction as non-core binding.

Reason: The literal human EIF3F–MYDGF (Q969H8) pair is supported by the HuRI two-hybrid screen and its retesting procedure. The retained record contains construct features: gal4 activation domain:c-c; gal4 dna binding domain:n-n. A current reference-product match does not prove fragment or isoform functional equivalence. Positive assay-context binding is retained; partner identity alone does not establish a specific EIF3F molecular function or native regulatory role, and the pair is not rejected from compartment expectations.

Summary: Retain the CCDC120 screen interaction as non-core binding.

Reason: The literal human EIF3F–CCDC120 (Q96HB5) pair is supported by the HuRI two-hybrid screen and its retesting procedure. The retained record contains construct features: gal4 activation domain:n-n; gal4 dna binding domain:n-n. A current reference-product match does not prove fragment or isoform functional equivalence. Positive assay-context binding is retained; partner identity alone does not establish a specific EIF3F molecular function or native regulatory role, and the pair is not rejected from compartment expectations.

Summary: Retain the PBX4 screen interaction as non-core binding.

Reason: The literal human EIF3F–PBX4 (Q9BYU1) pair is supported by the HuRI two-hybrid screen and its retesting procedure. The retained record contains construct features: gal4 activation domain:c-c; gal4 dna binding domain:n-n. A current reference-product match does not prove fragment or isoform functional equivalence. Positive assay-context binding is retained; partner identity alone does not establish a specific EIF3F molecular function or native regulatory role, and the pair is not rejected from compartment expectations.

Summary: Retain the MKRN2 screen interaction as non-core binding.

Reason: The literal human EIF3F–MKRN2 (Q9H000) pair is supported by the HuRI two-hybrid screen and its retesting procedure. The retained record contains construct features: gal4 activation domain:n-n; gal4 dna binding domain:n-n. A current reference-product match does not prove fragment or isoform functional equivalence. Positive assay-context binding is retained; partner identity alone does not establish a specific EIF3F molecular function or native regulatory role, and the pair is not rejected from compartment expectations.

Summary: Retain the MYOZ1 screen interaction as non-core binding.

Reason: The literal human EIF3F–MYOZ1 (Q9NP98) pair is supported by the HuRI two-hybrid screen and its retesting procedure. The retained record contains construct features: gal4 activation domain:n-n; gal4 dna binding domain:n-n. A current reference-product match does not prove fragment or isoform functional equivalence. Positive assay-context binding is retained; partner identity alone does not establish a specific EIF3F molecular function or native regulatory role, and the pair is not rejected from compartment expectations.

Summary: Retain the TIMM10B screen interaction as non-core binding.

Reason: The literal human EIF3F–TIMM10B (Q9Y5J6) pair is supported by the HuRI two-hybrid screen and its retesting procedure. The retained record contains construct features: gal4 activation domain:c-c; gal4 dna binding domain:n-n. A current reference-product match does not prove fragment or isoform functional equivalence. Positive assay-context binding is retained; partner identity alone does not establish a specific EIF3F molecular function or native regulatory role, and the pair is not rejected from compartment expectations.

Summary: Retain the DMWD screen interaction as non-core binding.

Reason: The literal human EIF3F–DMWD (G5E9A7) pair is supported by the reported neurodegeneration interaction screen and its pairwise two-hybrid retesting. The retained record contains construct features: sufficient binding region:316-649. A current reference-product match does not prove fragment or isoform functional equivalence. Positive assay-context binding is retained; partner identity alone does not establish a specific EIF3F molecular function or native regulatory role, and the pair is not rejected from compartment expectations.

Summary: Retain the DNALI1 screen interaction as non-core binding.

Reason: The literal human EIF3F–DNALI1 (O14645) pair is supported by the reported neurodegeneration interaction screen and its pairwise two-hybrid retesting. The retrieved pair record does not resolve an assayed insert sequence. A current reference-product match does not prove fragment or isoform functional equivalence. Positive assay-context binding is retained; partner identity alone does not establish a specific EIF3F molecular function or native regulatory role, and the pair is not rejected from compartment expectations.

Summary: Retain the KLF11 screen interaction as non-core binding.

Reason: The literal human EIF3F–KLF11 (O14901) pair is supported by the reported neurodegeneration interaction screen and its pairwise two-hybrid retesting. The retrieved pair record does not resolve an assayed insert sequence. A current reference-product match does not prove fragment or isoform functional equivalence. Positive assay-context binding is retained; partner identity alone does not establish a specific EIF3F molecular function or native regulatory role, and the pair is not rejected from compartment expectations.

Summary: Retain the HTRA2 screen interaction as non-core binding.

Reason: The literal human EIF3F–HTRA2 (O43464) pair is supported by the reported neurodegeneration interaction screen and its pairwise two-hybrid retesting. The retrieved pair record does not resolve an assayed insert sequence. A current reference-product match does not prove fragment or isoform functional equivalence. Positive assay-context binding is retained; partner identity alone does not establish a specific EIF3F molecular function or native regulatory role, and the pair is not rejected from compartment expectations.

Summary: Retain the KIF1B screen interaction as non-core binding.

Reason: The literal human EIF3F–KIF1B (O60333-2) pair is supported by the reported neurodegeneration interaction screen and its pairwise two-hybrid retesting. The retained record contains construct features: sufficient binding region:1192-1770. A current reference-product match does not prove fragment or isoform functional equivalence. Positive assay-context binding is retained; partner identity alone does not establish a specific EIF3F molecular function or native regulatory role, and the pair is not rejected from compartment expectations.

Summary: Retain the PRPF40A screen interaction as non-core binding.

Reason: The literal human EIF3F–PRPF40A (O75400-2) pair is supported by the reported neurodegeneration interaction screen and its pairwise two-hybrid retesting. The retained record contains construct features: sufficient binding region:716-930. A current reference-product match does not prove fragment or isoform functional equivalence. Positive assay-context binding is retained; partner identity alone does not establish a specific EIF3F molecular function or native regulatory role, and the pair is not rejected from compartment expectations.

Summary: Retain the WFS1 screen interaction as non-core binding.

Reason: The literal human EIF3F–WFS1 (O76024) pair is supported by the reported neurodegeneration interaction screen and its pairwise two-hybrid retesting. The retrieved pair record does not resolve an assayed insert sequence. A current reference-product match does not prove fragment or isoform functional equivalence. Positive assay-context binding is retained; partner identity alone does not establish a specific EIF3F molecular function or native regulatory role, and the pair is not rejected from compartment expectations.

Summary: Retain the HSPB1 screen interaction as non-core binding.

Reason: The literal human EIF3F–HSPB1 (P04792) pair is supported by the reported neurodegeneration interaction screen and its pairwise two-hybrid retesting. The retrieved pair record does not resolve an assayed insert sequence. A current reference-product match does not prove fragment or isoform functional equivalence. Positive assay-context binding is retained; partner identity alone does not establish a specific EIF3F molecular function or native regulatory role, and the pair is not rejected from compartment expectations.

Summary: Retain the LAMP2 screen interaction as non-core binding.

Reason: The literal human EIF3F–LAMP2 (P13473-2) pair is supported by the reported neurodegeneration interaction screen and its pairwise two-hybrid retesting. The retrieved pair record does not resolve an assayed insert sequence. A current reference-product match does not prove fragment or isoform functional equivalence. Positive assay-context binding is retained; partner identity alone does not establish a specific EIF3F molecular function or native regulatory role, and the pair is not rejected from compartment expectations.

Summary: Retain the NF2 screen interaction as non-core binding.

Reason: The literal human EIF3F–NF2 (P35240-4) pair is supported by the reported neurodegeneration interaction screen and its pairwise two-hybrid retesting. The retrieved pair record does not resolve an assayed insert sequence. A current reference-product match does not prove fragment or isoform functional equivalence. Positive assay-context binding is retained; partner identity alone does not establish a specific EIF3F molecular function or native regulatory role, and the pair is not rejected from compartment expectations.

Summary: Retain the BAG6 screen interaction as non-core binding.

Reason: The literal human EIF3F–BAG6 (P46379-2) pair is supported by the reported neurodegeneration interaction screen and its pairwise two-hybrid retesting. The retrieved pair record does not resolve an assayed insert sequence. A current reference-product match does not prove fragment or isoform functional equivalence. Positive assay-context binding is retained; partner identity alone does not establish a specific EIF3F molecular function or native regulatory role, and the pair is not rejected from compartment expectations.

Summary: Retain the ATXN1 screen interaction as non-core binding.

Reason: The literal human EIF3F–ATXN1 (P54253) pair is supported by the reported neurodegeneration interaction screen and its pairwise two-hybrid retesting. The retrieved pair record does not resolve an assayed insert sequence. A current reference-product match does not prove fragment or isoform functional equivalence. Positive assay-context binding is retained; partner identity alone does not establish a specific EIF3F molecular function or native regulatory role, and the pair is not rejected from compartment expectations.

Summary: Retain the CASP6 screen interaction as non-core binding.

Reason: The literal human EIF3F–CASP6 (P55212) pair is supported by the reported neurodegeneration interaction screen and its pairwise two-hybrid retesting. The retrieved pair record does not resolve an assayed insert sequence. A current reference-product match does not prove fragment or isoform functional equivalence. Positive assay-context binding is retained; partner identity alone does not establish a specific EIF3F molecular function or native regulatory role, and the pair is not rejected from compartment expectations.

Summary: Retain the PRPS1 screen interaction as non-core binding.

Reason: The literal human EIF3F–PRPS1 (P60891) pair is supported by the reported neurodegeneration interaction screen and its pairwise two-hybrid retesting. The retrieved pair record does not resolve an assayed insert sequence. A current reference-product match does not prove fragment or isoform functional equivalence. Positive assay-context binding is retained; partner identity alone does not establish a specific EIF3F molecular function or native regulatory role, and the pair is not rejected from compartment expectations.

Summary: Retain the CYCS screen interaction as non-core binding.

Reason: The literal human EIF3F–CYCS (P99999) pair is supported by the reported neurodegeneration interaction screen and its pairwise two-hybrid retesting. The retrieved pair record does not resolve an assayed insert sequence. A current reference-product match does not prove fragment or isoform functional equivalence. Positive assay-context binding is retained; partner identity alone does not establish a specific EIF3F molecular function or native regulatory role, and the pair is not rejected from compartment expectations.

Summary: Retain the TARDBP screen interaction as non-core binding.

Reason: The literal human EIF3F–TARDBP (Q13148) pair is supported by the reported neurodegeneration interaction screen and its pairwise two-hybrid retesting. The retained record contains construct features: sufficient binding region:1-260. A current reference-product match does not prove fragment or isoform functional equivalence. Positive assay-context binding is retained; partner identity alone does not establish a specific EIF3F molecular function or native regulatory role, and the pair is not rejected from compartment expectations.

Summary: Retain the SPRED1 screen interaction as non-core binding.

Reason: The literal human EIF3F–SPRED1 (Q7Z699) pair is supported by the reported neurodegeneration interaction screen and its pairwise two-hybrid retesting. The retrieved pair record does not resolve an assayed insert sequence. A current reference-product match does not prove fragment or isoform functional equivalence. Positive assay-context binding is retained; partner identity alone does not establish a specific EIF3F molecular function or native regulatory role, and the pair is not rejected from compartment expectations.

Summary: Retain the JPH3 screen interaction as non-core binding.

Reason: The literal human EIF3F–JPH3 (Q8WXH2) pair is supported by the reported neurodegeneration interaction screen and its pairwise two-hybrid retesting. The retrieved pair record does not resolve an assayed insert sequence. A current reference-product match does not prove fragment or isoform functional equivalence. Positive assay-context binding is retained; partner identity alone does not establish a specific EIF3F molecular function or native regulatory role, and the pair is not rejected from compartment expectations.

Summary: Retain the NUP58 screen interaction as non-core binding.

Reason: The literal human EIF3F–NUP58 (Q9BVL2) pair is supported by the reported neurodegeneration interaction screen and its pairwise two-hybrid retesting. The retrieved pair record does not resolve an assayed insert sequence. A current reference-product match does not prove fragment or isoform functional equivalence. Positive assay-context binding is retained; partner identity alone does not establish a specific EIF3F molecular function or native regulatory role, and the pair is not rejected from compartment expectations.

Summary: Retain the SH3GLB1 screen interaction as non-core binding.

Reason: The literal human EIF3F–SH3GLB1 (Q9Y371) pair is supported by the reported neurodegeneration interaction screen and its pairwise two-hybrid retesting. The retrieved pair record does not resolve an assayed insert sequence. A current reference-product match does not prove fragment or isoform functional equivalence. Positive assay-context binding is retained; partner identity alone does not establish a specific EIF3F molecular function or native regulatory role, and the pair is not rejected from compartment expectations.

Summary: Refine the established eIF3 assembly interaction to initiation-factor binding.

Reason: BioPlex 3.0 records the EIF3F–EIF3H AP-MS association. Native human F:H subcomplexes independently establish its initiation-assembly context; the screen does not establish a purified binary interface. GO:0031369 names binding to a polypeptide factor involved in translation initiation; it does not transfer the partner’s catalytic activity.

Proposed replacements: translation initiation factor binding

Summary: Refine the established eIF3 assembly interaction to initiation-factor binding.

Reason: BioPlex 3.0 records EIF3F–EIF3B association. Independent human eIF3 co-immunoprecipitation and complex architecture establish their shared initiation assembly without treating every AP-MS association as direct contact. GO:0031369 names binding to a polypeptide factor involved in translation initiation; it does not transfer the partner’s catalytic activity.

Proposed replacements: translation initiation factor binding

Summary: Refine the established eIF3 assembly interaction to initiation-factor binding.

Reason: BioPlex 3.0 records EIF3F–EIF3M association. Native human F:M subcomplexes provide independent mechanistic context for a specific initiation-factor-binding refinement. GO:0031369 names binding to a polypeptide factor involved in translation initiation; it does not transfer the partner’s catalytic activity.

Proposed replacements: translation initiation factor binding

Summary: Refine the established eIF3 assembly interaction to initiation-factor binding.

Reason: The released OpenCell table contains EIF3B as the endogenously tagged bait and EIF3F as a positive interactor (data row 8333). The MS identifier field contains O00303 and H0YDT6, so a unique splice form is not assigned. Established native eIF3 assembly supplies the functional context. GO:0031369 names binding to a polypeptide factor involved in translation initiation; it does not transfer the partner’s catalytic activity.

Proposed replacements: translation initiation factor binding

Summary: Refine the established eIF3 assembly interaction to initiation-factor binding.

Reason: The released U2OS AP-MS network contains the exact EIF3F–EIF3H edge 620520. Native human F:H subcomplexes independently support initiation-factor binding. The later correction concerns a loss-function formula, not this interaction edge. GO:0031369 names binding to a polypeptide factor involved in translation initiation; it does not transfer the partner’s catalytic activity.

Proposed replacements: translation initiation factor binding

Summary: Refine the established eIF3 assembly interaction to initiation-factor binding.

Reason: The released U2OS AP-MS network contains the exact EIF3F–EIF3B edge 620528. Independent human eIF3 assembly and co-immunoprecipitation support initiation-factor binding; the association is not relabeled as a binary assay. GO:0031369 names binding to a polypeptide factor involved in translation initiation; it does not transfer the partner’s catalytic activity.

Proposed replacements: translation initiation factor binding

Summary: Refine the established eIF3 assembly interaction to initiation-factor binding.

Reason: The released U2OS AP-MS network contains the exact EIF3F–EIF3M edge 620534. Native human F:M subcomplexes independently support this functional refinement. GO:0031369 names binding to a polypeptide factor involved in translation initiation; it does not transfer the partner’s catalytic activity.

Proposed replacements: translation initiation factor binding

Summary: Refine the supported cytoplasmic location to cytosol.

Reason: The broad UniRule/UniProt-location transfer is consistent with direct purification of native human eIF3 from HeLa cytosol. All eight captured Reactome source paths lead to human O00303 in a cytosolic eIF3 assembly. Cytosol includes protein complexes under the current definition; this refinement does not claim exclusive localization or deny context-specific nuclear redistribution.

Proposed replacements: cytosol

Summary: Retain the cytosolic localization in the curated initiation model.

Reason: The literal Reactome:R-HSA-156808 input/output and component chain reaches EIF3F entity R-HSA-72540, reference O00303, represented at residues 2–357 in cytosol. This is curated pathway evidence, corroborated by native human HeLa cytosol purification. No GTPase or other catalyst activity of a different component is transferred to EIF3F.

Summary: Retain the cytosolic localization in the curated initiation model.

Reason: The literal Reactome:R-HSA-156823 input/output and component chain reaches EIF3F entity R-HSA-72540, reference O00303, represented at residues 2–357 in cytosol. This is curated pathway evidence, corroborated by native human HeLa cytosol purification. No GTPase or other catalyst activity of a different component is transferred to EIF3F.

Summary: Retain the cytosolic localization in the curated initiation model.

Reason: The literal Reactome:R-HSA-157849 input/output and component chain reaches EIF3F entity R-HSA-72540, reference O00303, represented at residues 2–357 in cytosol. This is curated pathway evidence, corroborated by native human HeLa cytosol purification. No GTPase or other catalyst activity of a different component is transferred to EIF3F.

Summary: Retain the cytosolic localization in the curated initiation model.

Reason: The literal Reactome:R-HSA-72619 input/output and component chain reaches EIF3F entity R-HSA-72540, reference O00303, represented at residues 2–357 in cytosol. This is curated pathway evidence, corroborated by native human HeLa cytosol purification. No GTPase or other catalyst activity of a different component is transferred to EIF3F.

Summary: Retain the cytosolic localization in the curated initiation model.

Reason: The literal Reactome:R-HSA-72621 input/output and component chain reaches EIF3F entity R-HSA-72540, reference O00303, represented at residues 2–357 in cytosol. This is curated pathway evidence, corroborated by native human HeLa cytosol purification. No GTPase or other catalyst activity of a different component is transferred to EIF3F.

Summary: Retain the cytosolic localization in the curated initiation model.

Reason: The literal Reactome:R-HSA-72676 input/output and component chain reaches EIF3F entity R-HSA-72540, reference O00303, represented at residues 2–357 in cytosol. This is curated pathway evidence, corroborated by native human HeLa cytosol purification. No GTPase or other catalyst activity of a different component is transferred to EIF3F.

Summary: Retain the cytosolic localization in the curated initiation model.

Reason: The literal Reactome:R-HSA-72691 input/output and component chain reaches EIF3F entity R-HSA-72540, reference O00303, represented at residues 2–357 in cytosol. This is curated pathway evidence, corroborated by native human HeLa cytosol purification. No GTPase or other catalyst activity of a different component is transferred to EIF3F.

Summary: Retain the cytosolic localization in the curated initiation model.

Reason: The literal Reactome:R-HSA-72697 input/output and component chain reaches EIF3F entity R-HSA-72540, reference O00303, represented at residues 2–357 in cytosol. This is curated pathway evidence, corroborated by native human HeLa cytosol purification. No GTPase or other catalyst activity of a different component is transferred to EIF3F.

Summary: EIF3F is a constituent of the eIF3 complex.

Reason: The accessible source abstract identifies human HeLa eIF3 with all 13 subunits at stoichiometric abundance; the complete publisher Methods remain unavailable. Independent human native-complex and reconstitution evidence corroborates EIF3F membership.

Summary: EIF3F is a constituent of the eIF3 complex.

Reason: The human 11- and six-subunit recombinant eIF3 preparations include EIF3F. The native comparator is mouse, and failure to obtain sufficient F-minus material is not an activity measurement.

Summary: EIF3F is a constituent of the eIF3 complex.

Reason: Native human HeLa eIF3 mass spectrometry identifies EIF3F and resolves F:H:M-containing subcomplexes.

Summary: EIF3F is a constituent of the eIF3 complex.

Reason: Human 293T and HeLa eIF3 purification and RNA studies include EIF3F in the assembly; direct RNA crosslink assignments to other subunits are not transferred to F.

Summary: EIF3F is a constituent of the eIF3 complex.

Reason: The aggregate ARBA, mouse, InterPro and UniRule channels are genuine matches to the eIF3F product. Native human complex purification independently supports the predicted component membership.

Summary: EIF3F is a constituent of the eIF3 complex.

Reason: The source’s native human eIF3 composition supports the imported physical-association membership assertion; the unavailable complete Methods do not justify overturning that well-corroborated component assignment.

Summary: EIF3F is a constituent of the eIF3 complex.

Reason: The source identifies human p47 and its association with eIF3 through anti-p170 immunoprecipitation. Its distinct p66 RNA-binding result is not used as p47 evidence.

Summary: Specify cytoplasmic translational initiation.

Reason: The literal PTN000056988 IBD has human/mouse experimental grounding and a pombe cytoplasmic-initiation donor. The human product is inside that ancestral clade; neither donor count nor self-donation is a challenge. Human preinitiation-complex evidence makes the cytoplasmic process more informative than the broad parent.

Proposed replacements: cytoplasmic translational initiation

Summary: Specify cytoplasmic translational initiation.

Reason: The source identifies the native human eIF3 assembly; independent functional human reconstitution anchors the inferred initiation contribution. The supported context is cytoplasmic, without claiming every downstream translational outcome.

Proposed replacements: cytoplasmic translational initiation

Summary: Specify cytoplasmic translational initiation.

Reason: Native HeLa cytosolic eIF3 and its F:H:M module provide structural participation in the machinery. Functional human reconstitution corroborates the more precise cytoplasmic initiation process.

Proposed replacements: cytoplasmic translational initiation

Summary: Specify cytoplasmic translational initiation.

Reason: Human recombinant F-containing eIF3 supports 48S assembly. This is positive complex-level participation; the five-subunit negative control lacks C, not F, and is not represented as a matched F-minus assay.

Proposed replacements: cytoplasmic translational initiation

Summary: Specify cytoplasmic translational initiation.

Reason: ARBA00033914 genuinely matches the target FunFam plus Eutheria; the mouse and eIF3F-family channels are also retained. Human complex data support a cytoplasmic initiation refinement independently of the aggregate annotation.

Proposed replacements: cytoplasmic translational initiation

Summary: Refine broad peptidase activity to supported deubiquitinase activity.

Reason: IPR000555 genuinely maps the target MPN domain to peptidase activity, but domain naming alone cannot establish catalytic chemistry. The human Ub-GFP cleavage and associated cellular ubiquitination experiments independently support the more informative deubiquitinase function. This refinement does not assert a metal or cysteine catalytic mechanism.

Proposed replacements: deubiquitinase activity

Summary: The membrane-proteomics assertion remains unresolved.

Reason: The accessible source describes a YTS-cell membrane proteome that includes both likely membrane proteins and transiently associated proteins. The EIF3F target measurement and operative fractionation details could not be independently recovered after the preserved publisher failure. The term permits attached proteins and complexes, so neither lack of a transmembrane segment nor cytosolic eIF3 membership refutes it; preserve uncertainty rather than remove the experimental assertion.

Summary: Retain EIF3F membership in the 43S preinitiation complex.

Reason: The matching UniRule emits this component term. The captured Reactome 43S complex contains the eIF3 assembly and exact O00303 member, while human reconstitution supports the same initiation machinery. This names a transient assembly state and does not require exclusive residence.

Summary: Retain conserved initiation-factor binding within the eIF3 assembly.

Reason: The literal PTN000056988 IBD has a mouse experimental annotation for initiation-factor binding. The human leaf inherits that node, and independent native human F:H/F:M subcomplexes establish the functional assembly contact. The phylogenetic claim is not reduced to donor count or an unverified partner label.

Summary: Retain initiation-factor binding supported by conserved assembly.

Reason: The mouse protein mapping and eIF3F-family term output are verified. The mouse has an experimental initiation-factor-binding record, while native human F:H/F:M subcomplexes directly establish the same assembly role. No catalytic activity of a binding partner is transferred.

Summary: Retain the 48S preinitiation-complex component annotation.

Reason: The literal UniRule output agrees with human F-containing eIF3 preparations that support 48S toeprints and the captured Reactome component chains. This is complex membership during initiation, not an isolated F RNA-binding or catalytic assay.

Summary: Retain the self-interaction screen result without an obligate homodimer model.

Reason: Three literal HuRI records have O00303 on both interaction sides, supporting identical-protein binding in the assay. They are distinguished from all other target interaction records. No independently established physiological homodimer architecture, stoichiometry or autonomous function follows from this screen result.

Summary: Retain the orthology-supported synaptic context as non-core.

Reason: The transfer resolves to mouse Q9DCH4/ENSMUSP00000033342.7, whose SynGO EXP and IDA records both cite PMID:25211037 with the same UBERON:0002421 extension. These are two exported records from one source context, not independent studies. This is a credible curated localization transfer retained without claiming a directly inspected human synaptic assay or a dedicated synaptic function of EIF3F.

Summary: Retain membership in an eIF3 complex containing eIF3M.

Reason: The current term names an entire eIF3 complex containing the PCI protein eIF3M, not the isolated M subunit. The literal PTN000056988 IBD has mouse and pombe experimental grounding. Human native F:M/F:H:M complexes independently establish the matching architecture, so there is no target-specific reason to reject the conserved component claim.

Summary: Retain the conserved eIF3M-containing complex membership.

Reason: Mouse Q9DCH4 carries a direct experimental annotation for this complex, and its mapped protein identity is exact. Human native F:M and F:H:M subcomplexes independently support the corresponding eIF3M-containing assembly. The term does not assert that EIF3F is EIF3M.

Summary: Retain the viral IRES initiation context as a non-core role of eIF3.

Reason: The original uses rabbit-reticulocyte eIF3 and assigns UV-crosslinked bands principally by mobility, including a p47 band; it is not an isolated human EIF3F RNA-binding assay. Whole eIF3 promotes viral IRES initiation, and later native human eIF3 binds HCV IRES while retaining the F:H module. Together these support the contextual complex contribution without assigning direct RNA-contact chemistry to F or claiming a human-specific assay in the older source.

Summary: Retain the broader, experimentally supported deubiquitinase activity.

Reason: ARBA00026566 genuinely matches the target FunFam rather than merely the broad MPN fold. Human EIF3F-containing bacterial preparations cleave Ub-GFP, with distinct mouse active-probe and cellular rescue evidence. The broader DUB term accommodates this positive evidence while catalytic nucleophile, metal requirement and purified human kinetic constants remain unresolved.

Summary: Preserve inherited DUB activity while leaving metal dependence unresolved.

Reason: The human leaf inherits the literal PTN000056988 metal-DUB IBD, with genuine human PMID:21124883 and fly eIF3f1 PMID:38279019 experimental donors. The fly source has positive purified-protein Ub-rhodamine cleavage, distinct from cellular dTak1 ubiquitination. Neither inspected original establishes metal requirement by chelation/rescue, titration or ligand mutation; buffer MgCl2 and JAMM naming do not settle that mechanism. Generalize to DUB activity while preserving the positive node placement, not declaring inactivity or a wrong donor.

Proposed replacements: deubiquitinase activity

Summary: Generalize the positive human DUB assay without claiming a metal mechanism.

Reason: The original directly reports cleavage by human GST-full-length EIF3F and an MPN fragment in bacterial extracts; mouse active-probe, mutant and rescue experiments are separate. UniProt curates activated NOTCH1 deubiquitination and a requirement for DTX1 association from PMID:21124883. In that study, partial depletion of endogenous human EIF3F increased ubiquitinated murine activated Notch constructs, with rescue by murine wild-type but not the MPN mutant; DTX1 enabled the observed Notch-eIF3f association. These are positive cellular findings supporting the curated Notch context. The authors leave the precise monoubiquitinated Notch substrate form unresolved and retain an associated-DUB caveat. The inspected Methods and figure panels support DUB activity but contain no purified human metal-dependence experiment. GO:0101005 retains the activity while leaving its metal or thiol mechanism unassigned.

Proposed replacements: deubiquitinase activity