FANCG

Summary: The verified ancestral assertion supports FANCG participation in a broad DNA-damage response.

Reason: The actual target tree/MSA and IBD at PTN001433280 are valid. The mouse donor is reviewed Q9EQR6, MGI:1926471, with experimental IMP evidence from PMID:11719385 and PMID:11823446. Selected original Methods, Results and Figure 5 caption of PMID:11719385 show Fancg disruption, crosslink-induced chromosome damage and reduced survival, plus modest ionizing-radiation responses and deficient damage-induced Fancd2 monoubiquitination in mouse cells. PMID:11823446 positively reports MMC sensitivity but no X-ray, UV or MMS sensitivity in its MEFs; those cell/model-specific results do not erase the separately observed radiation response. Retain the broad term alongside the existing ICL-repair context rather than losing supported response breadth. This is pathway participation, not FANCG catalysis or a donor-count argument.

Summary: The verified inherited annotation agrees with direct human FA nuclear-complex membership.

Reason: The actual PANTHER v19 target path carries the GO:0043240 IBD at PTN001433280, grounded in human UniProtKB:O15287 itself. That self-WITH is legitimate experimental ancestry, not circularity. Human FANCG immunoprecipitation and fractionation in PMID:11063725 and nuclear purification/reciprocal immunoprecipitation in PMID:22343915 directly support complex membership. These assays do not make every assembly contact a purified binary interaction or assign autonomous ligase chemistry to FANCG.

Summary: FANCG has an established nuclear pool with more than one supported nuclear subcompartment.

Reason: The actual UniProt SL-0191 mapping supports nucleus. Human lymphoblast fractionation and FANCG immunoprecipitation in PMID:11063725, Figures 5–6, place FANCG in nuclear complexes; PMID:22343915, Figure 5H and its Results, additionally reports FANCG in the chromatin fraction after damage. HPA supports nuclear speckles. Retain nucleus rather than replacing all these pools by nucleoplasm, whose definition excludes chromosomes. UniProt major/minor-form wording does not establish natural splice products.

Summary: A positive cytoplasmic FANCG pool is retained as a contextual location.

Reason: UniProt SL-0086 maps to cytoplasm, and PMID:11063725 Figure 6 directly detects FANCG in nuclear and cytoplasmic fractions of human cells. PMID:17060495 additionally examines tagged human FANCG in HeLa cells and corrected lymphoblasts, with mitochondrial imaging and fractionation controls. Therefore cytoplasm cannot universally be replaced by soluble cytosol: it encompasses organelles as well. Retain this nonnuclear pool as non-core in this broad location context without calling it inactive, artifactual, or an established natural splice product. The inherited cell-cycle statement is background cited by the 2006 paper, not a newly re-executed assay.

Summary: The ARBA DNA-damage-response assertion is biologically supported, with a rule-version limitation retained.

Reason: The current official ARBA00026271 record outputs GO:0006974 and has 213 alternative condition sets. Its exact historical application to O15287 was not reconstructed: the target has CATH superfamily 1.25.40.10, but that alone does not prove membership in the rule’s narrower FunFam 1.25.40.10:FF:000750. Acceptance rests on independently verified FANCG damage-response biology, including human nuclear-complex evidence and mouse crosslink/radiation experiments, rather than a fabricated current rule match. The breadth exceeds ICL repair alone under the source-specific assay limits described for PMID:11719385 and PMID:11823446.

Summary: FANCG participates in ICL repair as a functional FA-core subunit.

Reason: The actual InterPro IPR039684 FANCG family entry maps to GO:0036297, and its PTHR15254 match covers displayed human O15287 residues 6–621. Human nuclear assembly experiments establish that FANCG supplies part of the repair machinery, while donor Fancg disruption produces crosslink-sensitive cells and deficient downstream Fancd2 activation. The inherited human-XRCC9/CHO complementation quotation is positive but the original was read only at abstract and selected Discussion scope. The inherited PMID:19965384 quotation concerns Xenopus ID2 depletion/add-back and pathway repair, not a purified human FANCG incision assay. Retain the specific repair process without transferring FANCL ligase, nuclease or polymerase chemistry to FANCG.

Summary: The FANCG-specific InterPro model supports FA nuclear-complex membership.

Reason: The official IPR039684 family entry explicitly maps to GO:0043240 and matches human O15287 over residues 6–621. This is a FANCG-family assertion, not an inference that every tetratricopeptide-repeat protein belongs to the FA core. Direct human FANCG nuclear-complex recovery in PMID:11063725 and PMID:22343915 corroborates membership. The inherited PMID:22266823 quotation is pathway context; its direct FAAP20–FANCA interaction is not reassigned to FANCG.

Summary: The FANCG–FANCA context supports the established integral complex-scaffold role.

Reason: The complete PMID:10627486 abstract explicitly reports a strong full-length FANCA/FANCG yeast two-hybrid interaction. Its complete operative text was unavailable; no unobserved controls are inferred. Human assembly and bridge evidence in PMID:11063725, PMID:12649160 and PMID:15262960 supports GO:0140378 protein complex scaffold activity. The repair-complex mechanism and current adaptor/scaffold distinction are detailed in the source guide.

Proposed replacements: protein complex scaffold activity

Summary: The FANCG–FANCA context supports the established integral complex-scaffold role.

Reason: The PMID:10652215 abstract reports strong FANCA/FANCG two-hybrid interaction and reduced binding by FANCG terminal fragments. Full Methods remain inaccessible, so exact fragment/assay details beyond that abstract are not invented. Human assembly and bridge evidence in PMID:11063725, PMID:12649160 and PMID:15262960 supports GO:0140378 protein complex scaffold activity. The repair-complex mechanism and current adaptor/scaffold distinction are detailed in the source guide.

Proposed replacements: protein complex scaffold activity

Summary: The FANCG–FANCA context supports the established integral complex-scaffold role.

Reason: PMID:11063725 human lymphoblast co-immunoprecipitation and fractionation show FANCA/FANCG association in the nuclear complex and restoration after FANCF complementation. This is assembled-complex evidence rather than an isolated measurement of every interface. Human assembly and bridge evidence in PMID:11063725, PMID:12649160 and PMID:15262960 supports GO:0140378 protein complex scaffold activity. The repair-complex mechanism and current adaptor/scaffold distinction are detailed in the source guide.

Proposed replacements: protein complex scaffold activity

Summary: The FANCG–FANCA context supports the established integral complex-scaffold role.

Reason: The positive PMID:12649160 abstract maps FANCG interaction to the FANCA amino terminus and explicitly identifies FANCG as the three-hybrid bridge connecting FANCA and FANCF. Its complete Methods were not recovered; the abstract is not presented as a new full-paper reading. Human assembly and bridge evidence in PMID:11063725, PMID:12649160 and PMID:15262960 supports GO:0140378 protein complex scaffold activity. The repair-complex mechanism and current adaptor/scaffold distinction are detailed in the source guide.

Proposed replacements: protein complex scaffold activity

Summary: The FANCG–FANCA context supports the established integral complex-scaffold role.

Reason: The actual PMID:16189514 supplementary Table S2 record at line 6513 identifies FANCA/FANCG as positive Y2H and literature-core overlap. The target co-affinity-purification cells are empty: global co-AP validation rates are not assigned to this pair. The read supplementary Methods describe reporter and autoactivation controls. Human assembly and bridge evidence in PMID:11063725, PMID:12649160 and PMID:15262960 supports GO:0140378 protein complex scaffold activity. The repair-complex mechanism and current adaptor/scaffold distinction are detailed in the source guide.

Proposed replacements: protein complex scaffold activity

Summary: The FANCG–FANCA context supports the established integral complex-scaffold role.

Reason: The source-specific IntAct record supports the curated FANCA/FANCG association, while the accessible PMID:17289582 abstract principally tests FANCM/FAAP24. Its full original was not recovered. Independently read human A/G/F assembly experiments substantiate the functional replacement; no FAAP24 DNA binding or FANCM chemistry is transferred to FANCG. Human assembly and bridge evidence in PMID:11063725, PMID:12649160 and PMID:15262960 supports GO:0140378 protein complex scaffold activity. The repair-complex mechanism and current adaptor/scaffold distinction are detailed in the source guide.

Proposed replacements: protein complex scaffold activity

Summary: The FANCG–FANCA context supports the established integral complex-scaffold role.

Reason: PMID:17396147 shows human FANCA/FANCG-containing nuclear fractions and complex association. Its negative three-hybrid tests for FAAP100 with FANCA/FANCG and the separately resolved FANCB–FANCL–FAAP100 module are retained. The evidence does not put FANCG into the catalytic BL100 unit. Human assembly and bridge evidence in PMID:11063725, PMID:12649160 and PMID:15262960 supports GO:0140378 protein complex scaffold activity. The repair-complex mechanism and current adaptor/scaffold distinction are detailed in the source guide.

Proposed replacements: protein complex scaffold activity

Summary: FANCG recognizes the SH3 domain of alpha-II spectrin.

Reason: The complete PMID:19102630 abstract explicitly reports target-specific yeast two-hybrid mapping to the SPTAN1 SH3 domain and confirmation by FANCG motif mutagenesis; FANCC and FANCF provide negative comparisons. GO:0017124 directly captures this domain-selective association, more precisely than generic protein binding or spectrin binding. Complete Methods and figure pixels were not recovered, so no additional affinity, native stoichiometry or construct detail is claimed. The proposed role in spectrin stabilization is distinguished from the demonstrated domain binding.

Proposed replacements: SH3 domain binding

Summary: FANCG associates with the viral transcriptional coregulator Tax.

Reason: PMID:22458338 reports the FANCG–Tax interaction in a pairwise-retested yeast two-hybrid pipeline with multiple reporter criteria, followed by Tax/LTR reporter assays in HEK293T cells and three FANCG shRNAs in Jurkat-LTR-Luc cells with viability normalization. This is more than an isolated screen hit. Tax is a transcriptional coregulator, and GO:0001221 captures binding to that partner in this experimentally supported viral-transcription context. The reporter phenotype does not itself demonstrate direct binding or identify the interface; pair-specific MAPPIT validation and a general host-transcription function are not claimed. The inherited supporting_text is the paper title, retained under the project quotation-preservation instruction; it is not the experimental support for this refinement. The supporting experiments are the pairwise-retest Results/Figure 1, Tax reporter Results/Figures 2–3 and FANCG-knockdown Methods, with public and cached locators in FANCG-source-guide.md.

Proposed replacements: transcription coregulator binding

Summary: The FANCG–FANCA context supports the established integral complex-scaffold role.

Reason: The explicit BioPlex 2.0 directed table records FANCG bait and FANCA prey (line 45975). The read Methods use human ORFeome C-terminal HA-FLAG constructs and 293T AP-MS. This is positive co-complex association, not a purified binary-interface assay. Human assembly and bridge evidence in PMID:11063725, PMID:12649160 and PMID:15262960 supports GO:0140378 protein complex scaffold activity. The repair-complex mechanism and current adaptor/scaffold distinction are detailed in the source guide.

Proposed replacements: protein complex scaffold activity

Summary: The FANCG–FANCA context supports the established integral complex-scaffold role.

Reason: PMID:31467278 reference-pair records include split-luciferase and MAPPIT tests. The literal curated partner is FANCA isoform O15360-3, but its experimental feature is sufficient-binding region 1–297. That fragment result is not a demonstration that full natural isoform 3 assembles or restores the core. The functional refinement uses independent full-complex evidence for FANCG and retains the original literal partner context. Human assembly and bridge evidence in PMID:11063725, PMID:12649160 and PMID:15262960 supports GO:0140378 protein complex scaffold activity. The repair-complex mechanism and current adaptor/scaffold distinction are detailed in the source guide.

Proposed replacements: protein complex scaffold activity

Summary: The FANCG–SUOX screen interaction is retained as non-core binding.

Reason: HuRI reports the FANCG–SUOX pair in array, pooling and validated-two-hybrid records, with study-level pairwise retesting and sequence confirmation. These are positive binary-association data, not independent native-cell validations. SUOX is an enzyme, but this screen does not establish an additional FANCG function in sulfite metabolism or the mechanism of its association with native SUOX. Retain the imported interaction as non-core and withdraw the partner-class replacement.

Summary: The FANCG–FANCA context supports the established integral complex-scaffold role.

Reason: The NeuroNet FANCA/FANCG record is an explicit mammalian luminescence-based interaction assay, distinct from the other FANCG yeast-screen contexts in this paper. The read STAR Methods define the tagged mammalian assay. Positive association corroborates the independently established repair-complex scaffold function; it does not imply all NeuroNet partners join that complex. Human assembly and bridge evidence in PMID:11063725, PMID:12649160 and PMID:15262960 supports GO:0140378 protein complex scaffold activity. The repair-complex mechanism and current adaptor/scaffold distinction are detailed in the source guide.

Proposed replacements: protein complex scaffold activity

Summary: The FANCG–FANCA context supports the established integral complex-scaffold role.

Reason: BioPlex 3.0 contains FANCG/FANCA in both 293T and HCT116 networks and in the corresponding directed G-bait/A-prey records. Sequence-verified tagged ORFeome AP-MS and gene-level peptide grouping are explicit in the Methods. The result corroborates co-complex association without establishing identical behavior of every natural product. Human assembly and bridge evidence in PMID:11063725, PMID:12649160 and PMID:15262960 supports GO:0140378 protein complex scaffold activity. The repair-complex mechanism and current adaptor/scaffold distinction are detailed in the source guide.

Proposed replacements: protein complex scaffold activity

Summary: The FANCG–FANCA context supports the established integral complex-scaffold role.

Reason: The exact PMID:37398436 reference-pair records contain BRET and luminescence co-precipitation measurements, with tagged human proteins in HEK293 cells. The read main describes the reference-set benchmark and LuTHy controls; it does not make this pair a viral interaction or establish a new full-length natural-product assignment. The current preprint source is retained as supplied. Human assembly and bridge evidence in PMID:11063725, PMID:12649160 and PMID:15262960 supports GO:0140378 protein complex scaffold activity. The repair-complex mechanism and current adaptor/scaffold distinction are detailed in the source guide.

Proposed replacements: protein complex scaffold activity

Summary: The FANCG–FANCA context supports the established integral complex-scaffold role.

Reason: The actual PMID:40205054 AP-MS network contains the FANCG/FANCA edge in human U2OS cells. The read Methods specify sequence-validated C-terminal HA-FLAG ORFeome constructs and CompPASS filtering. The saved edge is undirected, so no bait/prey direction is inferred from endpoint ordering. Association supports the same independently demonstrated structural role, not intrinsic enzymatic activity. Human assembly and bridge evidence in PMID:11063725, PMID:12649160 and PMID:15262960 supports GO:0140378 protein complex scaffold activity. The repair-complex mechanism and current adaptor/scaffold distinction are detailed in the source guide.

Proposed replacements: protein complex scaffold activity

Summary: The inherited authored proposal now specifies integral protein-complex scaffold activity.

Reason: The positive original IPI rationale and quotations remain attached to their historical sources. Actual human A/G/F complex experiments and the explicit FANCG bridge reported by PMID:12649160 support FANCG holding an integral repair assembly together. Current GO:0030674 specifically directs protein-complex scaffolds to GO:0140378. This retains structural biology while correcting the ontology subclass. The old blanket statement that FANCG has no DNA-binding activity is not reused: source-specific damaged-DNA binding remains unresolved. No intrinsic ligase activity or specialized ligase–substrate-adaptor positioning is inferred. This is the explicit scientific successor of the prior authored proposal, not a GOA source edit. The current GO term describes the measured structural role; action NEW retains its machine-independent proposal provenance. This authored duplicate is now retired because the same scaffold assertion is already proposed as the replacement of the imported PMID:10627486 binding row and supported by the other assembly contexts. The row and its quotation associations remain historical provenance; retirement does not reject scaffold activity.

Summary: HPA supports nuclear-speck localization without establishing a speck-specific repair function.

Reason: The actual HPA ENSG00000221829 record assigns supported nuclear-speck localization using HPA006695 in HeLa, REH and U2OS. Its 138-residue antigen exactly matches displayed O15287 residues 123–260. Selected HeLa and U2OS images were viewed and show discrete extranucleolar nuclear puncta; REH pixels were not inspected. Retain the positive location as non-core because these images do not establish that the FA repair activity occurs within the speckles. Antigen identity and HPA validation do not prove absolute antibody specificity, and no independent colocalization quantification or natural-product-specific localization is claimed. The inherited supporting_text is the GO_REF method title and is retained as provenance only; it is not target-specific experimental support. The localization judgment rests on the HPA target record and inspected images described here and in the source guide. The quotation stays unchanged under the project preservation instruction; its presence does not make the method title evidence for this protein.

Summary: Human FANCG is experimentally detected in the chromatin-associated FA core.

Reason: PMID:22343915 Figure 5H explicitly includes FANCG in the chromatin fraction, with H2A as loading control and untreated/MMC/HU conditions. The accompanying Results distinguish FANCA/FANCG loading from FAAP20-dependent repair signaling. The actual figure and chromatin-fractionation Methods were inspected. Retain the curated localization without turning chromatin association into isolated damaged-DNA recognition or autonomous ligase activity. The inherited abstract excerpt concerns FAAP20 and FANCA, not a FANCG-localization experiment. It remains unchanged as inherited provenance under the project preservation instruction and is not the basis of this ACCEPT. The evidence is the original full-text FANCG result and Figure 5H, accessed independently of the abstract-only cache; direct article and image links are in FANCG-source-guide.md.

Summary: FANCG contributes structurally to the machinery of interstrand cross-link repair.

Reason: PMID:19965384 directly executes Xenopus FANCI–FANCD2 depletion/add-back in a replication-coupled cisplatin-ICL repair system. That pathway experiment is not an isolated human FANCG assay. Human FANCG-containing nuclear-core assembly and functional complementation in independently read PMID:11063725/15262960/9256465 supply the target connection. Retain the NAS process because FANCG performs structural work in the repair machinery, without transferring incision, polymerase or ubiquitin-transfer catalysis.

Summary: The NAS Fanconi-complex membership agrees with positive human purification evidence.

Reason: The separately retained NAS context from PMID:22343915 is supported by the same demonstrated FANCG-containing nuclear complex as its IDA context. It remains a distinct imported assertion, not an additional independent experimental replicate. Neither context assigns catalytic transfer to FANCG.

Summary: Reactome supports a cytosolic FANCG pool in the PKR-response assembly.

Reason: Actual R-HSA-9835411 input/output routes through FA-core:HSP70 assemblies reach human FANCG entity 420788, residues 1–622, whose own compartment is GO:0005829 cytosol. The event emphasizes FANCC-dependent binding and inhibition of PKR. The complete accessible PMID:15299030 abstract reports FANCA/FANCC/FANCG functional interactions in human bone-marrow and FA-cell systems, but the operative full Methods were not accessed here. Retain the positive modeled cytosolic pool as contextual non-core localization; neither the model nor this abstract proves purified binary FANCG–PKR binding or intrinsic FANCG kinase/inhibitor chemistry.

Summary: Reactome places FANCG in nucleoplasm during FA-core assembly.

Reason: Reactome:R-HSA-6785126 was traced through actual participant links to human O15287 entity 419532, residues 1–622, with its own GO:0005654 compartment. FANCG is a direct input and part of the assembled output. The summation explicitly describes nucleoplasmic assembly before loading onto damaged DNA. Human nuclear-complex fractionation and purification in PMID:11063725 and PMID:22343915 independently corroborate a nuclear assembly pool. These ten modeled contexts are not ten independent localization experiments, and no untested alternative product or partner chemistry is inferred.

Summary: Reactome places FANCG in nucleoplasm during ID2 and UBE2T recruitment.

Reason: Reactome:R-HSA-6785342 was traced through actual participant links to human O15287 entity 419532, residues 1–622, with its own GO:0005654 compartment. FANCG remains in the core-containing input and output. The named FANCL–UBE2T recruitment role does not make FANCG an E2 enzyme. Human nuclear-complex fractionation and purification in PMID:11063725 and PMID:22343915 independently corroborate a nuclear assembly pool. These ten modeled contexts are not ten independent localization experiments, and no untested alternative product or partner chemistry is inferred.

Summary: Reactome places FANCG in nucleoplasm during ID2 monoubiquitination.

Reason: Reactome:R-HSA-6785361 was traced through actual participant links to human O15287 entity 419532, residues 1–622, with its own GO:0005654 compartment. Input, output and composite-catalyst routes contain FANCG. FANCL and UBE2T perform ubiquitin transfer; this location annotation does not assert autonomous FANCG ligase activity. Human nuclear-complex fractionation and purification in PMID:11063725 and PMID:22343915 independently corroborate a nuclear assembly pool. These ten modeled contexts are not ten independent localization experiments, and no untested alternative product or partner chemistry is inferred.

Summary: Reactome places FANCG in nucleoplasm during nuclease recruitment.

Reason: Reactome:R-HSA-6785732 was traced through actual participant links to human O15287 entity 419532, residues 1–622, with its own GO:0005654 compartment. The target persists in core-containing input and output; the recruited nucleases and their docking proteins have separate roles. Human nuclear-complex fractionation and purification in PMID:11063725 and PMID:22343915 independently corroborate a nuclear assembly pool. These ten modeled contexts are not ten independent localization experiments, and no untested alternative product or partner chemistry is inferred.

Summary: Reactome places FANCG in nucleoplasm during ICL unhooking.

Reason: Reactome:R-HSA-6785986 was traced through actual participant links to human O15287 entity 419532, residues 1–622, with its own GO:0005654 compartment. The target occurs in input, output and composite-catalyst routes. The named nucleases perform incision, not every component of the core-containing assembly. Human nuclear-complex fractionation and purification in PMID:11063725 and PMID:22343915 independently corroborate a nuclear assembly pool. These ten modeled contexts are not ten independent localization experiments, and no untested alternative product or partner chemistry is inferred.

Summary: Reactome places FANCG in nucleoplasm during POLN recruitment.

Reason: Reactome:R-HSA-6786155 was traced through actual participant links to human O15287 entity 419532, residues 1–622, with its own GO:0005654 compartment. FANCG occurs in core-containing input and output; POLN is the recruited DNA polymerase. Human nuclear-complex fractionation and purification in PMID:11063725 and PMID:22343915 independently corroborate a nuclear assembly pool. These ten modeled contexts are not ten independent localization experiments, and no untested alternative product or partner chemistry is inferred.

Summary: Reactome places FANCG in nucleoplasm during POLN translesion synthesis.

Reason: Reactome:R-HSA-6786166 was traced through actual participant links to human O15287 entity 419532, residues 1–622, with its own GO:0005654 compartment. Input, output and composite-catalyst routes contain FANCG. POLN performs synthesis; the event does not establish FANCG polymerase chemistry. Human nuclear-complex fractionation and purification in PMID:11063725 and PMID:22343915 independently corroborate a nuclear assembly pool. These ten modeled contexts are not ten independent localization experiments, and no untested alternative product or partner chemistry is inferred.

Summary: Reactome places FANCG in nucleoplasm during ID2 deubiquitination.

Reason: Reactome:R-HSA-6786171 was traced through actual participant links to human O15287 entity 419532, residues 1–622, with its own GO:0005654 compartment. FANCG remains in the substrate assembly and output. The separate USP1:WDR48 complex performs deubiquitination. Human nuclear-complex fractionation and purification in PMID:11063725 and PMID:22343915 independently corroborate a nuclear assembly pool. These ten modeled contexts are not ten independent localization experiments, and no untested alternative product or partner chemistry is inferred.

Summary: Reactome places FANCG in nucleoplasm during ATR–ATRIP recruitment.

Reason: Reactome:R-HSA-6788385 was traced through actual participant links to human O15287 entity 419532, residues 1–622, with its own GO:0005654 compartment. The target persists in input and output. Named RPA, FANCM, FAAP24 and FANCL contacts do not establish a binary FANCG–ATRIP interaction. Human nuclear-complex fractionation and purification in PMID:11063725 and PMID:22343915 independently corroborate a nuclear assembly pool. These ten modeled contexts are not ten independent localization experiments, and no untested alternative product or partner chemistry is inferred.

Summary: Reactome places FANCG in nucleoplasm during ATR-dependent phosphorylation.

Reason: Reactome:R-HSA-6788392 was traced through actual participant links to human O15287 entity 419532, residues 1–622, with its own GO:0005654 compartment. The target is present in input, output and composite-catalyst routes. ATR is the kinase and the named substrates are other components; kinase activity is not transferred to FANCG. Human nuclear-complex fractionation and purification in PMID:11063725 and PMID:22343915 independently corroborate a nuclear assembly pool. These ten modeled contexts are not ten independent localization experiments, and no untested alternative product or partner chemistry is inferred.

Summary: FANCG is a constituent of the human Fanconi core discussed and assayed in the FAAP20 study.

Reason: The PMID:22266823 study supports core-complex association in human cells, while its binary assay distinguishes direct FAAP20–FANCA binding from negative FAAP20–FANCG binding. Retain the curator-supported complex membership, independently corroborated by actual FANCG purification/co-IP in PMID:22343915 and PMID:20347428. A negative specific pair does not negate membership, and no direct G–FAAP20 interface is invented.

Summary: FANCG is positively identified in the purified human Fanconi nuclear core.

Reason: PMID:22343915 actual human HeLa nuclear-complex purification and reciprocal co-immunoprecipitation support FANCG membership. The paper distinguishes FANCA-mediated FAAP20 association and stability from the FANCG-containing assembly. Retain this IDA membership without assigning FAAP20 ubiquitin recognition or ligase catalysis to FANCG.

Summary: FANCG is a component of the human FA core recovered with FAAP20.

Reason: PMID:22705371 explicitly identifies FANC-G after human HeLa nuclear-extract fractionation/FANCA immunoprecipitation and mass spectrometry. Figure 1 and Results also show anti-FANCG recovery of FAAP20 and co-elution of A/G/FAAP20. Recruitment experiments centered on FANCA/FANCD2 and separate chicken genetics are not relabeled as individual human FANCG assays. The demonstrated membership is sound.

Summary: Human FANCG is directly identified in the nuclear Fanconi core complex.

Reason: PMID:20347428 Results and Figure 1 describe HeLa nuclear-extract gel filtration and FANCM immunoprecipitation followed by mass spectrometry/immunoblotting, explicitly naming FANC-G among recovered core proteins. The separate observation that most FANCM/MHF resides outside the core limits generalization to all FANCM assemblies but does not weaken the positive FANCG membership. No FANCM ATPase or MHF DNA-remodeling function is transferred.

Summary: FANCG binds the enzyme PRDX3 and can protect it from proteolytic cleavage.

Reason: PMID:17060495 shows target-specific human FANCG/PRDX3 two-hybrid interaction and reciprocal co-IP of tagged FANCG with endogenous PRDX3, with diminished interaction by FANCG G546R. Actual Figure 3F additionally shows in-vitro-translated FANCG protecting translated PRDX3 from rat calpain II cleavage. GO:0019899 captures binding to the established peroxiredoxin enzyme without requiring a catalytic-regulation claim. The experiment does not distinguish shielding a bound substrate from direct calpain inhibition, and mitochondrial-lysate peroxidase activity is not intrinsic FANCG peroxidase catalysis. No more specific PRDX3-binding MF was established in the current term search.

Proposed replacements: enzyme binding

Summary: A mitochondrial FANCG pool is experimentally supported.

Reason: PMID:17060495 provides actual human FANCG/mitochondrial-marker imaging and marker-controlled fractionation, including tagged HeLa cells and stably complemented FA-G lymphoblasts. Figure 2G shows similar localization of the PRDX3-binding-defective G546R mutant; Figure 2H uses corrected cells and is not relabeled endogenous wild-type imaging. Retain the positive contextual organellar location alongside the established nuclear pool, without making every FANCG molecule mitochondrial or inventing a natural splice-product assignment.

Summary: FANCG contributes to mitochondrial morphology and PRDX3-associated oxidative protection in the tested cells.

Reason: PMID:17060495 reports abnormal mitochondrial morphology in primary FA-G fibroblasts and partial rescue after FANCG reintroduction; actual Figure 4 preserves the incomplete rescue. GO:0007005 includes mitochondrial morphogenesis and organization, so the positive phenotype fits this contextual process. Separate Figure 3F demonstrates protection of PRDX3 from calpain cleavage, strengthening a plausible performed protective contribution, but does not prove that this reaction alone causes every morphology change or that FANCG is a peroxidase. Retain as non-core without treating partial rescue as no effect.

Summary: The specific damaged-DNA-binding claim remains unresolved.

Reason: The complete accessible PMID:9806548 abstract identifies human FANCG/XRCC9 by FA-G complementation and patient mutations, but does not expose a damaged-DNA-binding assay. The publisher full-text attempt did not recover the operative paper. GO:0003684 requires binding damaged DNA; repair complementation, chromatin recruitment and membership in a DNA-associated complex do not alone resolve that molecular activity. Preserve the TAS assertion as uncertain rather than infer absence of binding, wrong-gene attribution, or a contradictory experimental result.

Summary: FANCG contributes to DNA repair with experimentally supported breadth beyond one damage assay.

Reason: The original PMID:9256465 Methods and Results show human XRCC9/FANCG complementation in Chinese hamster UV40: partial resistance rescue for MMC/cisplatin and other genotoxins and strong correction of chromosomal aberrations, with unchanged sister-chromatid exchange. Independent human repair-complex assembly establishes FANCG as a structural participant rather than damaged DNA serving only as substrate. Retain broad DNA repair because the source scope is not exhausted by ICL alone. This does not assign nuclease chemistry or an isolated damaged-DNA-binding assay.

Summary: The inherited human homologous-recombination proposal retains positive conserved support but needs bounded reassessment.

Reason: PMID:12861027 directly demonstrates reduced SCneo homologous recombination in chicken DT40 FANCG mutants and rescue by chicken FANCG; human FANCG did not complement in that host. This is positive evidence for conserved involvement and the host-specific failure is not evidence of human functional loss. PMID:18212739 additionally contains human/hamster FANCG-dependent BRCA2–FANCD2–XRCC3 complex association and separate chicken epistasis, correcting the old claim that both papers only use chicken. That human structural association is relevant indirect mechanistic support, but the precise human HR participation represented by this authored proposal remains uncertain in the inspected evidence. The original positive ISS rationale and both quotations remain visible; direct human endpoint assays are not a universal prerequisite for GO inference. Same-role human FA-core comparator records lack this term at the bounded cut, which is a caution rather than an exclusion rule. A retained ISS interpretation based on the conserved reporter phenotype plus human structural mechanism remains a defensible alternative for final synthesis; this proposal is not biological negation. The specific authored human HR assertion is withdrawn from active proposals while its positive conserved evidence and unresolved review remain preserved. This is curator withdrawal of our proposal, not withdrawal by GO, loss of an imported GOA annotation, or biological negation.

Summary: The FANCG–FANCF context supports the established integral complex-scaffold role.

Reason: PMID:11063725 shows endogenous human FANCF/FANCG coassociation and FANCF-dependent restoration of FANCA–FANCG assembly. Its reticulocyte pairwise FANCF/FANCG experiment was negative under those conditions; this remains distinct from positive cellular coassociation and subsequent interface mapping. Human assembly and bridge evidence in PMID:11063725, PMID:12649160 and PMID:15262960 supports GO:0140378 protein complex scaffold activity. The repair-complex mechanism and current adaptor/scaffold distinction are detailed in the source guide.

Proposed replacements: protein complex scaffold activity

Summary: The FANCG–FANCF context supports the established integral complex-scaffold role.

Reason: The positive PMID:12649160 abstract maps FANCG binding to the FANCF carboxyl terminus and reports loss with FANCG Leu71Pro. Human FANCF deletion assays in the independently read PMID:15262960 show graded loss of FANCG association; some weak-co-IP mutants still complement MMC sensitivity, so binding strength is not equated with an all-or-none functional threshold. Human assembly and bridge evidence in PMID:11063725, PMID:12649160 and PMID:15262960 supports GO:0140378 protein complex scaffold activity. The repair-complex mechanism and current adaptor/scaffold distinction are detailed in the source guide.

Proposed replacements: protein complex scaffold activity

Summary: The FANCG–ZNF329 HuRI association is retained as contextual binding.

Reason: The exact official HuRI target records and matching IntAct source entries identify this human pair in the yeast two-hybrid pipeline. Array/pooling/validated records are stages of that pipeline, not three independent physiological validations. Main text reports retesting and sequence confirmation; global MAPPIT/GPCA benchmarks are not pair-specific evidence. The partner record describes only a possible transcriptional-regulatory role; a zinc-finger family label alone is insufficient to assign a specific transcription-factor-binding activity to FANCG. No scaffold or repair role for this pair was established. Retain the positive association as non-core under the supported-binding policy.

Summary: The FANCG–TPRX1 screen interaction is retained as non-core binding.

Reason: The HuRI records identify FANCG with the literal partner TPRX1 isoform Q8N7U7-2 in the controlled two-hybrid pipeline. TPRX1 is a DNA-binding transcription factor, but the tested association does not establish a FANCG transcriptional complex or activity; historical clone labels do not prove full natural-product equivalence. Retain the positive interaction as non-core rather than specializing it solely from the partner classification.

Summary: The literal FANCG–TCEANC isoform context remains a supported screen association.

Reason: The exact official HuRI target records and matching IntAct source entries identify this human pair in the yeast two-hybrid pipeline. Array/pooling/validated records are stages of that pipeline, not three independent physiological validations. Main text reports retesting and sequence confirmation; global MAPPIT/GPCA benchmarks are not pair-specific evidence. Current Q8N8B7-2 and historical Q8N8B7-1/ORF856 labels are retained without silently remapping products. The TFIIS-like family/domain description does not establish the exact transcriptional activity or an informative FANCG regulator function in this assay. Retain non-core binding; no RNA-polymerase or elongation activity is transferred to FANCG.

Summary: The FANCG–CCHCR1 association is supported in the HuRI construct context.

Reason: The exact official HuRI target records and matching IntAct source entries identify this human pair in the yeast two-hybrid pipeline. Array/pooling/validated records are stages of that pipeline, not three independent physiological validations. Main text reports retesting and sequence confirmation; global MAPPIT/GPCA benchmarks are not pair-specific evidence. Current Q8TD31-3 and historical Q8TD31-1/ORF53053 labels remain distinct. No CCHCR1-specific binding MF, core-scaffold role, or native centrosomal/skin process follows from this pair alone. Retain positive non-core binding without generalizing across products.

Summary: The FANCG–PRPF18 HuRI interaction is retained as non-core binding.

Reason: The exact official HuRI target records and matching IntAct source entries identify this human pair in the yeast two-hybrid pipeline. Array/pooling/validated records are stages of that pipeline, not three independent physiological validations. Main text reports retesting and sequence confirmation; global MAPPIT/GPCA benchmarks are not pair-specific evidence. The exact historical PRPF18 ORF961 and current Q99633 context are retained. PRPF18 spliceosomal participation does not establish FANCG spliceosome activity or RNA binding. Specificity assessment did not establish a more informative supported FANCG MF for this pair.

Summary: The FANCG–DMWD screen context supports contextual non-core binding.

Reason: The actual NeuroNet NN1.0 target row and matching IntAct records support this human-protein pair in the yeast two-hybrid pipeline. The read STAR Methods address autoactivation and reporter selection. These records do not by themselves establish a native neuronal role, repair-complex membership or target-specific mammalian validation. The literal G5E9A7 construct is annotated as residues 316–649. A WD-repeat fragment association does not establish a specific scaffold function, enzyme class or a myotonic-dystrophy mechanism for FANCG.

Summary: The FANCG–KLF11 screen interaction is retained as non-core binding.

Reason: The NeuroNet table and matched IntAct array, pooling and validated-two-hybrid records support FANCG–KLF11 association. KLF11 is a DNA-binding transcription factor, but no target-specific transcriptional complex or functional assay for this pair was recovered. Retain the curated binding as non-core without adding a transcription-factor-binding specialization from the partner label alone.

Summary: The FANCG–KIF1B screen interaction is retained as non-core binding.

Reason: The NeuroNet association is with the sufficient-binding region 1192–1770 of KIF1B isoform O60333-2, not a demonstrated full motor–FANCG complex. The inspected GO:0019894 definition additionally requires stoichiometric interaction; the reporter assay does not establish stoichiometry. Preserve the positive fragment interaction as non-core without a kinesin-binding replacement or transport-function inference.

Summary: The FANCG–WFS1 screen context supports contextual non-core binding.

Reason: The actual NeuroNet NN1.0 target row and matching IntAct records support this human-protein pair in the yeast two-hybrid pipeline. The read STAR Methods address autoactivation and reporter selection. These records do not by themselves establish a native neuronal role, repair-complex membership or target-specific mammalian validation. Wolframin is an ER membrane/stress-response protein, but the pair does not identify a specific FANCG enzyme, chaperone or receptor-regulator activity. Native membrane topology and ER stress participation are not inferred.

Summary: The FANCG–HSPB1 screen interaction is retained as non-core binding.

Reason: The NeuroNet target record and validated-two-hybrid curation support FANCG–HSPB1 association. They do not establish a chaperone-dependent FANCG activity, an effect on HSPB1 function or an independently validated native complex. Preserve this positive association as non-core; the Hsp27-binding label alone does not supply that missing functional interpretation.

Summary: The FANCG–VIM screen interaction is retained as non-core binding.

Reason: NeuroNet supports a controlled binary association of FANCG with vimentin. This does not assay binding to an assembled intermediate filament or demonstrate a FANCG cytoskeletal role. Retain the imported protein-pair interaction as non-core and withdraw the classification-based cytoskeletal-protein-binding replacement.

Summary: The FANCG–CYP3A4 screen interaction is retained as non-core binding.

Reason: The NeuroNet two-hybrid records support FANCG–CYP3A4 association under the screening conditions. They do not establish a native membrane-associated complex or FANCG involvement in CYP3A4 catalysis. CYP3A4 enzyme identity is not evidence for an additional FANCG enzymatic role; preserve the positive imported interaction as non-core without the partner-class replacement.

Summary: The FANCG–GFAP screen interaction is retained as non-core binding.

Reason: The NeuroNet array, pooling and validated-two-hybrid records support FANCG–GFAP association. They do not assay an assembled GFAP filament or establish a glial cytoskeletal function for FANCG. Retain the positive protein-pair evidence as non-core and withdraw the classification-based cytoskeletal-protein-binding replacement.

Summary: The FANCG–PECAM1 screen context supports contextual non-core binding.

Reason: The actual NeuroNet NN1.0 target row and matching IntAct records support this human-protein pair in the yeast two-hybrid pipeline. The read STAR Methods address autoactivation and reporter selection. These records do not by themselves establish a native neuronal role, repair-complex membership or target-specific mammalian validation. A PECAM1 association does not establish a physiological adhesion interface, receptor agonism or vascular function of FANCG. The available target assay does not justify replacing this with cell-adhesion activity merely from the partner name.

Summary: The FANCG–NDUFV2 screen context supports contextual non-core binding.

Reason: The actual NeuroNet NN1.0 target row and matching IntAct records support this human-protein pair in the yeast two-hybrid pipeline. The read STAR Methods address autoactivation and reporter selection. These records do not by themselves establish a native neuronal role, repair-complex membership or target-specific mammalian validation. NDUFV2 is an electron-transfer subunit of respiratory Complex I; the enzyme activity belongs to the multicomponent complex. The individual-pair record does not establish binding to an intact enzyme complex or FANCG respiratory chemistry.

Summary: The FANCG–RAB5A screen interaction is retained as non-core binding.

Reason: NeuroNet supports FANCG–RAB5A association in its controlled yeast two-hybrid pipeline. The records do not distinguish GDP/GTP-dependent binding, identify an effector mechanism or establish a FANCG vesicle-trafficking function. Retain the imported association as non-core rather than adding a functional specialization solely from RAB5A identity.

Summary: The FANCG–NOS3 screen interaction is retained as non-core binding.

Reason: The NeuroNet records support the FANCG–NOS3 pair in the controlled two-hybrid pipeline. They do not establish an endogenous NOS3 complex, binding stoichiometry or FANCG regulation of nitric-oxide production. Retain the positive association as non-core and withdraw the enzyme-specific replacement rather than treating the screen as a new FANCG nitric-oxide pathway function.

Summary: The FANCG–ATXN3 screen interaction is retained as non-core binding.

Reason: NeuroNet reports FANCG association with an ATXN3 N-terminal sufficient-binding region whose endpoint is unspecified in the inspected IntAct record. The record does not establish equivalence to the complete active deubiquitinase or a FANCG role in its catalysis. Preserve the positive construct-specific interaction as non-core and withdraw the ubiquitin-specific-protease-binding replacement.

Summary: The FANCG–PRPS1 screen interaction is retained as non-core binding.

Reason: The NeuroNet table and validated-two-hybrid records support FANCG–PRPS1 association. They do not establish FANCG engagement with the native PRPS1 catalytic assembly or a nucleotide-synthesis function. Retain this positive imported interaction as non-core without specializing it solely because PRPS1 is an enzyme.

Summary: The FANCG–PLD1 screen interaction is retained as non-core binding.

Reason: NeuroNet supports FANCG–PLD1 association in the yeast two-hybrid pipeline. The inspected records do not establish a native phospholipase complex, lipid-dependent association or an effect on PLD1 turnover. Preserve the imported association as non-core and withdraw the partner-class replacement; no phospholipase activity is assigned to FANCG.

Summary: The FANCG–GLE1 screen context supports contextual non-core binding.

Reason: The actual NeuroNet NN1.0 target row and matching IntAct records support this human-protein pair in the yeast two-hybrid pipeline. The read STAR Methods address autoactivation and reporter selection. These records do not by themselves establish a native neuronal role, repair-complex membership or target-specific mammalian validation. GLE1 participates in RNA export and helicase regulation, but the pair does not show FANCG acting on RNA, regulating a helicase or forming a defined transport scaffold. No more informative supported FANCG activity was established.

Summary: The FANCG–SPRED1 screen context supports contextual non-core binding.

Reason: The actual NeuroNet NN1.0 target row and matching IntAct records support this human-protein pair in the yeast two-hybrid pipeline. The read STAR Methods address autoactivation and reporter selection. These records do not by themselves establish a native neuronal role, repair-complex membership or target-specific mammalian validation. SPRED1 is a signaling regulator rather than the small GTPase itself. This pair cannot be relabeled small-GTPase binding, Ras inhibition or a specific scaffold activity without the corresponding mechanism.

Summary: The FANCG–JPH3 screen context supports contextual non-core binding.

Reason: The actual NeuroNet NN1.0 target row and matching IntAct records support this human-protein pair in the yeast two-hybrid pipeline. The read STAR Methods address autoactivation and reporter selection. These records do not by themselves establish a native neuronal role, repair-complex membership or target-specific mammalian validation. Junctophilin-3 membrane-junction roles do not demonstrate that FANCG tethers membranes or regulates a channel. Retain the positive association without importing the partner’s localization or neurological phenotype.

Summary: The FANCG–APBB2 screen context supports contextual non-core binding.

Reason: The actual NeuroNet NN1.0 target row and matching IntAct records support this human-protein pair in the yeast two-hybrid pipeline. The read STAR Methods address autoactivation and reporter selection. These records do not by themselves establish a native neuronal role, repair-complex membership or target-specific mammalian validation. The literal APBB2 isoform-2 association is preserved. Its multidomain adaptor roles do not establish a particular FANCG transcription-coregulator or APP-processing activity in the inspected assay; no equivalence to other APBB2 products is assumed.

Summary: The FANCG–NUP58 screen context supports contextual non-core binding.

Reason: The actual NeuroNet NN1.0 target row and matching IntAct records support this human-protein pair in the yeast two-hybrid pipeline. The read STAR Methods address autoactivation and reporter selection. These records do not by themselves establish a native neuronal role, repair-complex membership or target-specific mammalian validation. NUP58 nuclear-pore membership is a property of the partner. The pair does not show FANCG to be a pore subunit, transport receptor or channel regulator; no specific activity replacement was established.