The verified source3 seed is human AIFM1, UniProt O95831, HGNC:8768: 77 PENDING annotation objects, 29 reference records (21 PMIDs), and six alternative products. The coordinator checked current main d35dcc30b44924f79c0510b281ae824aa536848a, canonical/alias directories, the intended branch and seven separate open-PR title queries. AIFM1 and aliases PDCD8, NAMSD, AUNX1, AIF, CMTX4 and DFNX5 had no overlapping canonical directories or open reviews. /tmp/AIFM1-root-baseline.json and tmp/aifm1-current-preflight.json record the seed and remote checks; /tmp/AIFM1-seed-snapshot.json retains every parsed source object before review. Source identity, qualifiers, reference identifiers and six alternative-product records are immutable.
Default Falcon research with the prescribed Perplexity-lite fallback and 1200-second limit was launched concurrently with normal publication caching. The first offline launch could not resolve an absent tool dependency (/tmp/AIFM1-provider-attempt.log). One corrected online launch using supported per-process temporary UV directories failed DNS while retrieving the research-client dependency for both providers (/tmp/AIFM1-provider-online-attempt.log, terminal exit 1). No provider report was produced or manufactured. The parallel normal fetch-gene-pmids command confirmed all 21 seeded PMIDs already cached (exit 0; /tmp/AIFM1-fetch-seeded.log). Existing caches were not overwritten. Subsequent work is manual primary-source review.
A local gocams/index.tsv search for O95831/AIFM1 and the mouse/rat donor accessions returned no entries. That absence is not evidence for a missing annotation. Live QuickGO donor-annotation API requests for Q9JM53 and Q9Z0X1 were inaccessible through the web tool; the initial UniProt web pages returned JavaScript fallback text. Exact source tracing will be distinguished from target-level biological support.
PMID:39806100(https://pmc.ncbi.nlm.nih.gov/articles/PMC11832770/) is fully cached and was read through the human complex structure, cellular interaction/localization, substrate binding and refolding experiments. The crystal uses an engineered AIF-W196A/CHCHD4 N-terminal fusion, with solution, mutational and cellular corroboration; it is not an unmodified membrane-embedded holocomplex. Figure 6 independently tests NADH alone, unactivated AIF, NADH-activated AIF and competing CHCHD4-N45 peptide. Activated AIF increases CHCHD4 substrate binding and CHCHD6 oxidative folding. AIF therefore performs scaffold/regulatory work in the disulfide-relay import pathway; CHCHD4, with ALR regeneration, performs the disulfide chemistry. The H2O2 readout in Figure 5 reflects electron flow through ALR and must not be quoted as direct AIF H2O2 production.
Crucially, the same study finds CHCHD4 interaction-defective and N-terminal deletion variants can still localize to mitochondria in U2OS cells (Figure 2), whereas disrupting its own disulfide import motif impairs import. This refines the earlier interpretation of AIF as a direct carrier importing CHCHD4: AIF supports the relay through activation of an existing CHCHD4 chaperone, which can then fold/import nascent CHCHD4 in trans. The broader import participation remains directly supported.
PMID:26004228(https://pubmed.ncbi.nlm.nih.gov/26004228/) is locally abstract-only. The original author manuscript was recovered from University of Glasgow, 55 PDF pages including supplementary material. Its Results identify endogenous AIF–CHCHD4 association in human cervical cancer cells, recombinant binding, and NADH/NADPH-enhanced interaction. Depletion experiments in human U2OS cells and mouse stem-cell rescue are separate scopes. The exact respiratory-chain assembly assertion requires distinguishing AIF's demonstrated import-machinery role from downstream respiratory-complex loss. Reading continues; no assembly action is decided merely from the abstract.
PMID:15775970(https://pmc.ncbi.nlm.nih.gov/articles/PMC1142539/) is fully cached. HeLa endogenous protein, tagged constructs, differential digitonin permeabilization, proteinase protection and carbonate extraction resolve inner-membrane anchoring with the bulk C-terminal domain in the intermembrane space and a short N-terminal portion facing matrix. Mature protein is processed at M53/A54; death-associated release requires a separate cleavage near residue 101. Figure 5 compares full-length and cleavage-resistant constructs, with endogenous confirmation in Figure 8. Cytosolic/nuclear redistribution after actinomycin D remains detectable with zVAD-fmk. Matrix exposure of the small N terminus does not establish a matrix catalytic location.
PMID:33168626(https://pmc.ncbi.nlm.nih.gov/articles/PMC7948984/) is fully cached. HeLa staurosporine and actinomycin-D assays examine PAR-bearing 89-kDa PARP1 fragments, AIF co-immunoprecipitation and nuclear redistribution. PARP inhibition and cleavage/catalytic PARP1 mutants distinguish PAR-dependent binding from an unqualified direct binary AIF–PARP1 interaction. This is a caspase-activated apoptosis route with AIF participation; it must not be called universally caspase-independent parthanatos. AIF is not assigned PARP1's polymerase or DNA-cleaving chemistry.
Live AmiGO defines GO:0003954 by NADH-to-acceptor reduction, not proton pumping or obligatory respiratory complex I membership. GO:0016174 specifically requires oxygen reduction to H2O2. The latter is narrower chemistry, not a synonym for any NADH oxidoreductase assay. GO:0003954 is a child of GO:0016651 (acting on NAD(P)H).
Live GO:0043495 covers binding that brings a protein/complex to a membrane to establish or maintain localization. GO:0160203 describes oxidative-folding-driven import of small cysteine-containing proteins across the outer membrane via TOM and is a child of GO:0045041. AIF's experimentally resolved CHCHD4 anchor/activation is active participation, rather than inference from loss-of-function necessity alone. These terms are already seeded; no NEW process is needed.
PMID:23217327(https://pubmed.ncbi.nlm.nih.gov/23217327/) was read in the original University of Milan author PDF, independently recovered by the peer reviewer. Table 1 measures recombinant human WT/E493V AIF reduction of DCIP, ferricyanide and cytochrome c. Figure 3C follows reoxidation of NADH-reduced flavin by atmospheric oxygen at 452 nm; this does not identify H2O2 as a measured product. Figure 3E compares human WT/mutant DNA gel retardation. Patient muscle and fibroblast results address nuclear localization and apoptosis; this particular allele does not show the same respiratory deficit as every other AIFM1 variant. The local cache is abstract-only, and was not overwritten. Human acceptor-reduction chemistry is independently represented by the exact curated reaction in the seeded UniProt record: Reaction=A + NADH + H(+) = AH2 + NAD(+);.
PMID:27178839(https://pubmed.ncbi.nlm.nih.gov/27178839/) explicitly studies human AIF V243L, G262S, G308E and G338E. Its abstract and accessible publisher introduction support FAD/NADH redox properties; the full peroxide assay was not recovered by either reviewer. Thus the IDA peroxide-specific row and its PAINT/TAS counterparts remain UNDECIDED. This is an assay-access/product-specificity limit, not evidence that AIF cannot generate peroxide. The more general NADH dehydrogenase chemistry is positively supported. PMID:24914854(https://pubmed.ncbi.nlm.nih.gov/24914854/) explicitly identifies human protein, coenzyme/FAD occupancy, solution stoichiometry and a nondimerizing variant; it supports ligand binding and redox-linked dimerization without settling the physiological electron acceptor.
The original PMID:17094969(https://pubmed.ncbi.nlm.nih.gov/17094969/) abstract reports binding and functional translation inhibition competitively relieved by EIF3G, including a coupled transcription/translation system and cellular observations. This is sufficient positive evidence to refine generic binding to GO:0030371 translation repressor activity, whose definition concerns antagonizing ribosome-mediated translation and does not require RNA binding. It remains a mature-AIF/death-associated context, not a separate ubiquitous basal-translation core. Original confocal images were not recovered; the source-linked nuclear and perinuclear localization in UniProt is retained with that limitation.
Additional primary interaction source recovery used indexed original PMC text. PMID:21364629, PMC3032331 Figure 6 and co-IP Methods show AIF association with WT and LEA-deleted PRELI in human-cell lysates, while only WT PRELI protects mitochondria and prevents AIF nuclear redistribution. PMID:22371500, PMC3339961 Figures 2-5 show reciprocal HeLa co-IP, GST-SIP/in-vitro-translated AIF binding, mitochondrial retention and U2OS perturbation. These establish interaction/regulation, not a new intrinsic AIF binding function. Six generic protein-binding rows are removed on the uninformative-term policy, without denying the associations. The TSC22D4 network pair and the BioPlex KANK2 supplement entry were not independently recovered; GOA/UniProt partner identity remains preserved. The influenza paper explicitly identifies AIFM1 among human PA-associated proteins, so no wrong-gene assertion is made.
PMID:21467298(https://pubmed.ncbi.nlm.nih.gov/21467298/), original PMC3086524, was read through indexed full Results/Methods. Purified AIF and PAR-binding-region variants are mouse proteins, including the AIF constructs expressed in HeLa. PAR binding is tested with appropriate positive/negative controls; the mutants preserve other measured properties but lose PAR-triggered release/death. Mouse neuronal rescue provides additional context. Human-cell evidence in cached PMID:33168626 corroborates PAR-dependent behavior, but does not convert the mouse direct binding experiment to human IDA. The mouse source annotation is traced through the MGI Aifm1 annotation page.
The live GO:0070266 necroptotic process definition emphasizes RIPK/MLKL machinery; parthanatos is a related synonym, not an exact equivalence. Historical AIF/PAR studies do not resolve that mechanistic boundary. The existing positive-necroptosis ISS remains UNDECIDED rather than being removed or reclassified from a title. No claim is made that every regulated necrotic death is canonical necroptosis.
PMID:18056262(https://pubmed.ncbi.nlm.nih.gov/18056262/) was read in the author-uploaded original paper. Methods specifies full-length human AIF (the source's 1-612 numbering; GenBank AF100928) and an N-terminal truncation. Thapsigargin/Scythe loss-rescue experiments use primary mouse embryonic fibroblasts, with human-cell expression/co-IP steps. This supports a bounded ER-stress death contribution, with human protein and mouse host distinguished; it does not place AIF constitutively in the ER. The old 612-residue construct description is reported as stated, not silently reconciled to the current 613-residue UniProt sequence.
PMID:19418225(https://pubmed.ncbi.nlm.nih.gov/19418225/) was verified from its primary PubMed abstract: glucose oxidase supplies sustained H2O2, AIF knockdown affects death and nuclear redistribution accompanies pyknosis. Human UniProt explicitly attributes the target H2O2-response function to this source. The cell-line identity was not inferred from MeSH or the paper title. The response is retained as non-core with the exact rat donor chain unresolved; it is not peroxide synthesis by AIF.
Rat UniProt Q9JM53/Ensembl ENSRNOP00000008503 identity is independently supported by NCBI Gene 83533. Specific paper chains for the transferred ischemia, toxic-substance, neuron-apoptosis, estradiol, hypoxia, nitric-oxide, glutamate and aldosterone terms could not be recovered from the available annotation interfaces. They remain UNDECIDED. Orthology is not rejected, and no biological divergence or donor error is asserted merely because a source could not be traced.
PMID:16365034(https://pubmed.ncbi.nlm.nih.gov/16365034/) explicitly places AIFsh (C-terminal 353-613) in cytosol before nuclear translocation after HeLa overexpression. PMID:16644725(https://pubmed.ncbi.nlm.nih.gov/16644725/) distinguishes mitochondrial redox-active AIFsh2 from nuclear-apoptogenic forms. PMID:20111043(https://pubmed.ncbi.nlm.nih.gov/20111043/) compares membrane retention of AIF1 and alternative-exon-2b AIF2 (UniProt O95831-3). All six seeded alternative-product objects are preserved, as are the lack of isoform qualifiers on these original annotation rows; biological scope is explained in reasons rather than rewriting source assertions.
PMID:39979311(https://pubmed.ncbi.nlm.nih.gov/39979311/), original PMC11842818, was inspected through indexed Results/Figures 5-6 and Methods. Human postmortem tissue and a human cancer-cell splice model are separate from purified mouse AIF/AIF3 protein experiments. The paper reports GSSG reduction by purified mouse AIF, comparison with AIF3 and cellular GSR/AIF perturbations. This is a useful physiological-substrate question; it is not automatically a new human glutathione-reductase core. The exon-2/3-skipping product called AIF3 in this paper is not UniProt O95831-3, which is the AIF2 alternative-exon product. Construct nomenclature was not merged. No NEW redox or process annotation is proposed from this study.
The full 2015 CHCHD4 paper supports respiratory-chain biogenesis and engineered import-bypass rescue. GO:0033108 is under protein-complex assembly and mitochondrial organization. Its aggregation/arrangement wording does not require direct structural-subunit contact by every participant: an assembly chaperone can participate. The experiments leave uncertain whether AIF performs an assembly step beyond supplying import machinery. Therefore the assembly IMP remains UNDECIDED, not a categorical exclusion based on absent co-precipitation. The independently resolved AIF anchor/activation work supports the already seeded import and disulfide-relay process terms. There is no new core process term or NEW annotation row, so no pathway-completeness gap is alleged from GO-CAM absence.
All 77 seeded annotation objects were manually assessed: 44 ACCEPT, 9 KEEP_AS_NON_CORE, 5 MODIFY, 6 generic-binding REMOVE and 13 UNDECIDED. The three cores distinguish NADH-acceptor chemistry, membrane anchoring/activation of CHCHD4, and DNA-binding nuclear death effector activity. All 29 original reference identities and six alternative products are preserved; five references were added (four primary studies plus the read UniProt record), for 34 total. No NEW rows were added. Broad mitochondrial assertions retain their source resolution; proteomic/HPA evidence is not promoted to matrix or membrane resolution. The partially extracted PMID:20833797 body retains full_text_unavailable: false, matching its true cache metadata, with extraction limitations in the reference review.
Normal fetch attempts for PMID:21467298, PMID:18056262 and PMID:39979311 ended with DNS failure, exit 1, zero records (/tmp/AIFM1-new-sources-fetch.log). A subsequent normal single-source attempt for PMID:19418225 also ended with DNS failure, exit 1, zero records (/tmp/AIFM1-hydrogen-peroxide-fetch.log). These four real missing cache gates remain explicit; external primary access does not replace machine-cache provenance. No source record was hand-authored and no provider report was invented. References used only as bibliography inside the immutable UniProt source are not recursively asserted as supporting literature. The final census includes authored YAML/notes and any genuine provider report; no provider report exists for this failed launch.
The annotation-reviewer consultation independently recovered and checked the full human PMID:23217327 biochemical study and bounded PMID:27178839 access, agreeing that generic NADH chemistry is established while peroxide-product specificity remains unresolved. This was a bounded source consultation, not an independent read of all 77 decisions. The coordinator is reviewing the stable full draft separately. Validation/history/render results will be recorded after terminal completion.
Full just validate human AIFM1 completed successfully (exit 0; /tmp/AIFM1-validation.log) with two warning categories: the four missing primary caches and a local ontology-label snapshot mismatch for GO:0160203. The validator expects mitochondrial disulfide relay system; live official AmiGO GO:0160203, checked again on 2026-09-27, gives protein import into the intermembrane space via the disulfide relay system as the current name and the shorter phrase as a synonym. The verified current label is retained in the authored core, and all machine-source labels remain untouched. DRAFT status preserves both limitations.
The coordinator independently read all 77 decisions and propagation assessments, all 34 references, the three cores and complete notes, and verified the key translation and import/assembly distinctions; no biological blocker was identified in that read. Final checks include source/ref/isoform preservation, case-sensitive whitespace-normalized cached quote matching, absence of YAML anchors/trailing spaces, history validation and HTML rendering. Current-main preflight at 30a9290881824baaacab2b681d7808f572c4cef6 confirmed the AIFM1 directory and all 13 required source3 auxiliary publication paths remain absent (tmp/aifm1-source3-current-main.json). Only those exact normal-cache bytes accompany this gene package; no InterPro machine auxiliaries are publication dependencies.
Final integrity checks passed for all 77 source objects, 29 original reference identities and six alternative products. All 109 supporting-text entries match their cache with case preserved and whitespace normalized; none passes only by case folding. YAML contains no anchors, aliases or trailing spaces. History validation, HTML rendering, and final schema/best-practices/GOA validation after the two exact UniProt core snippets all completed with exit 0. The final explicit citation census is 25 PMIDs, with 21 cached and the four access gates above unresolved; there are no genuine provider-report citations to add. /tmp/AIFM1-integrity-report.json and the final manifest preserve the detailed checks and exact file hashes.
The exact published head 0f1d705e859ef64aaea9d24d8ce9e04f29577e63 of PR #3295 and all five local gene files were checked before this follow-up. Source14 supplied four normal-fetch records without overwriting any canonical source: PMID:18056262, PMID:19418225, PMID:21467298 and PMID:39979311. The coordinator's import receipt is tmp/source14-canonical-import-receipt.json (SHA256 45e4e452e6eaad0c359bc0ff3d8ba2006bde464894cd5dd3cc2af5c073a5b74e). The earlier DNS failures above remain historical facts; all four required records are now present.
The official PMID:39979311 supplement was downloaded directly after the web viewer rejected its size. Read-only local PDF/text artifacts are /tmp/AIFM1-39979311-supplement.{pdf,txt}; they are not manufactured publication-cache records. Figure S6 and Table S4 were read for assay/construct scope. The original main article is also available at Nature Communications. The structured source-review receipt records the exact downloaded bytes and access route. No raw source bytes were edited.
All 77 annotation source objects and decisions, 34 reference identities, six alternative products, and three complete core objects remain unchanged. Only the four reference access/scope assessments, the two recovered-full-text flags, and the existing GR question are updated. The authored YAML/notes census remains 25 distinct PMIDs, now all cached; there is no genuine provider report to add and no new PMID dependency. The prior GO:0160203 label-snapshot advisory is independent of source availability; final validation will determine the remaining status warning.
Final source14 validation passed (just validate human AIFM1, exit 0) with one remaining GO:0160203 ontology-label snapshot advisory and no missing-reference warning. YAML remains DRAFT for that documented advisory; source closure permits normal PR review. History validation/render and the final integrity receipt cover unchanged 77 complete annotation objects, three core objects, 34 reference identities and six alternative products; all 109 cached quotes pass case-sensitive whitespace-normalized matching. The four exact new publication files are absent from the existing PR base and are included without alteration in the follow-up manifest.
The coordinator independently read the complete source14 semantic delta and appended notes, checked the original PMID:39979311 Methods/Results distinctions, and accepted the bounded clarification with all annotation/core judgments preserved. Live AmiGO GO:0160203 was rechecked on 2026-09-27 and still gives the longer authored name, confirming the remaining warning is a local ontology snapshot mismatch.
The exact current PR #3295 head f561e9e633f6c348e9a063b6c04599ca0ef01b55, formal review 5330345721 and full comment 5855866784 were checked before edits. All five canonical gene files and both published histories matched that head. The following changes preserve all 77 source assertions and actions, original reference identities, core terms and biological scope.
supported_by; the two papers remain contextual additional citations. Neither endocrine-response conservation nor neuronal injury-response conservation is inferred from general death-effector activity.supporting_text_fulltext field or converted into an authored-notes quote proxy. The nucleus row for this same source likewise retains its primary reference and external-access flag while dropping the unrelated redox quote; the original patient-muscle localization Results provide that source's scope.Two alternative-product display names contained an unambiguous parser leak from UniProt synonym evidence: 4 (AIFsh2 {ECO:0000303|PubMed:16644725}) becomes 4 (AIFsh2), and 6 (AIFsh3 {ECO:0000303|PubMed:16644725}) becomes 6 (AIFsh3). All six accessions, sequence notes, biological synonyms and source rows are unchanged; the raw UniProt evidence tags remain untouched in the source file. No source or previously published history was edited.
The DOI/PMC/PMID census still comprises the same 25 required PMIDs, all cached. There is no genuine provider report, hidden provider bibliography or new source dependency. The known GO:0160203 local-label advisory remains separate from source closure; final validation and exact file/source checks are recorded in the new session manifest.
The follow-up passed full gene validation with only the existing GO:0160203 label-snapshot warning. All 90 retained cached supporting quotations pass case-sensitive, whitespace-normalized substring checking; none requires case folding. The integrity check preserves all 77 source assertions/actions, all 34 complete reference assessments, the three core mechanisms and every product field except the two documented display names. The coordinator independently read all 14 changed annotation blocks and the core/product-name changes, checked the localization and PAR-binding experiments, and found no biological blocker. History validation and rendering also passed; source bytes and the two published histories remain unchanged.