This is a substantive audit of the existing COMPLETE human ACSL4 review, performed under the review, annotation-reviewer and core-function-synthesizer instructions. All 51 seeded annotation rows and all 37 references/findings were inspected. The machine-sourced annotation fields remain unchanged. No NEW annotations were added.
ACSL4 is the approved human symbol (HGNC:3571; NCBI Gene 2182; UniProt O60488). Historical symbols include ACS4, FACL4, LACS4, MRX63 and MRX68. The NCBI gene record supplies the HGNC identity; the ClinGen campaign lists the association as Definitive. Alias and symbol searches found no overlapping open ACSL4 review PR. The coordinator's attempted exhaustive open-PR path query timed out with HTTP 504, so an exhaustive repository-wide overlap check is not claimed.
The seven gene files matched authoritative GitHub main at e0d565bddebefd72f7c1c32562b588f75a893f69 before editing. The review blob was d2e60977b646187ae1b17058c0a636ddf61ccc15; the HTML blob was 5236844b28c3c723f5a53e2bb35b84b9ca33e8e9. Local snapshots and the publication manifest record the remaining source blobs. There was no pre-existing notes file.
The required default Falcon research launch (1200-second observation allowance, with perplexity-lite fallback) ran alongside publication caching. The first attempt failed when uvx tried to create its tool environment in a protected default directory. A supported per-process retry used UV_TOOL_DIR=/tmp/aigr-uv-tools, UV_TOOL_BIN_DIR=/tmp/aigr-uv-bin, and UV_CACHE_DIR=/tmp/aigr-uv-cache. Both provider invocations then failed before research because their dependency installation could not resolve PyPI. No fresh report was produced and no provider process remained live. The existing genuine Falcon, Cyberian and Perplexity reports remain unedited; they are research leads, not substitutes for primary evidence. just fetch-gene-pmids human ACSL4 reported all 26 PMIDs already cached; no publication or Reactome cache was changed.
The local cache is full text for PMID:29450800, PMID:34800366 and PMID:38720107. The other 23 PMID caches are abstract-only, including files with duplicated abstract material under a full-text-looking heading. Their full_text_unavailable: true flags reflect local cache availability even where a separate external full-text route was successfully read. External access is documented below rather than misrepresented as a cache update.
The human cDNA study PMID:9598324 directly establishes a functional arachidonate-preferring fatty acid-CoA ligase. The rat enzyme in PMID:9096315 was expressed in bacteria and purified; it preferred arachidonate/EPA and retained lower-affinity palmitate activity. That study must not be described as a purified-human experiment.
The full publisher article for PMID:31061331 (publisher full text) was read, including Methods, Fig. 2/product assays and Results. Human ACSL4 variants cloned from Caco-2 cells were expressed in Sf9 cells and partially purified. Both produced acyl-CoAs from AA/EPA (C20) and DHA/adrenate (C22). The cellular expression host does not change the species of the human enzyme. The same paper explicitly leaves membrane topology unresolved after detecting both variants in pellet and supernatant. It does not justify a universal integral-membrane/type-III claim.
Live GO definitions were checked rather than inferred from common lipid terminology:
The original full patient study PMID:11889465 was not recovered. The Siena repository lists restricted full text. Its accessible abstract reports reduced enzyme activity in patient lymphoblastoid cells but does not identify a >22-carbon substrate. The GO:0031957 IMP row is therefore UNDECIDED. Neither rejecting that activity nor accepting it from C22 assays would be justified. The independent broad enzyme activity and broad cytoplasm assignment remain compatible with positive human evidence and curator judgment.
The two previous core entries described overlapping broad/specific versions of one ligase activity. They are now one core, using GO:0004467 to cover the actual HUFA substrate range, linked to long-chain acyl-CoA production and positive regulation of ferroptosis. Existing arachidonate-specific annotations remain accepted. Broad keyword mappings for nucleotide binding and ligase activity are refined to ATP binding and long-chain fatty acid-CoA ligase; two broad lipid-metabolism rows are refined to long-chain fatty acid metabolism. Broad fatty-acid metabolism itself remains acceptable pathway coverage.
The IEA donor is mouse Acsl4 (UniProt Q9QUJ7; ENSMUSP00000033634), not an assumed paralog. The precise original donor experiment for the biosynthesis assertion was not recovered, and the corresponding propagation source status remains UNRESOLVED.
The target assertion nevertheless has positive human pathway support. Cached R-HSA-548843 explicitly names ACSL4 as the catalyst that activates arachidonate at the ER. Its parent R-HSA-75876 explicitly places ACSL activation before ELOVL condensation and includes the arachidonate activation step. ACSL4 is doing reaction chemistry within this pathway, not simply being an input or being consumed. Retain the existing GO:0006633 annotation, with a bounded activation-in-elongation rationale. Do not describe ACSL4 as a carbon-condensing fatty-acid synthase or create a NEW process annotation. The coordinator independently read both cached Reactome records and agreed with this distinction.
The pathway's prose uses a >20-carbon convention for “very long-chain,” unlike GO:0031957's >22 definition. Its name or membership cannot establish ligation of a >22-carbon free fatty acid. The direct C20 arachidonate activation event still supports the existing long-chain acyl-CoA biosynthesis annotation.
For PMID:24269233 (author manuscript), the Hokkaido University author manuscript was read. Results 3.1 identifies human ACSL4 among constructs restoring deficient sphingolipid-to-glycerophospholipid metabolism in yeast. Results 3.2 distinguishes human variant 1 mainly at the plasma membrane, with an ER pool, from variant 2 predominantly at the ER in HeLa cells. The manuscript also identifies the ACSL4 enzyme tested in the preceding PMID:22633490 work. The latter's ALDH3A2-focused title is not evidence of an ACSL4 misattribution. ACSL3-specific droplet imaging in this manuscript is not reassigned to ACSL4.
The cached full PMID:29450800 (PMC full text) distinguishes endogenous ACSL4's calnexin-associated ER distribution from ACSL3 trafficking compartments using HT1080/MCF-7 fractionation and confocal microscopy. This is stronger ER-specific evidence than the original general cloning quote or provider location summary.
The full PMID:23455425 primary Nature article was recovered as an external transcription, matched by title/authors/DOI. Fig. 1c and adjacent Results identify HEK293 MAM fractions using FACL4 and calnexin. This is the direct FACL4/ACSL4 basis for retaining ER–mitochondria contact-site localization. The local abstract alone does not contain the result. No autophagy-process annotation is inferred from use as a fraction marker.
In cached full PMID:38720107, mitochondrial subfractionation/immunoelectron microscopy and imaging support outer-membrane ACSL4 with additional intermembrane-space signal. The review quotes the actual ACSL4 localization result instead of a generic statement about PCK2. The same study directly measures AA-d8-CoA formation by purified human-cell ACSL4 and tests activity-dependent ferroptosis susceptibility. PCK2 is the kinase; ACSL4 is its ligase substrate.
The MitoCoP HTP row retains mitochondrion resolution. Although full PMID:34800366 is cached, the ACSL4-specific supplementary identification was not independently extracted. Independent human-cell localization supports the compatible curated assignment, but does not turn the HTP source into an outer-membrane experiment. Likewise, the YTS membrane proteome (PMID:19946888) remains membrane, without arbitrary ER refinement or integral-topology inference.
The original HuH7 droplet proteome (PMID:14741744) is abstract-only and prominently describes ACSL3. The same authors' later primary JLR article, as exposed in indexed Discussion text, explicitly compares greater droplet ACSL3 abundance with ACSL4 and cites the original study. This corroborates the compatible curated ACSL4 IDA location; it does not mean the original full identification was recovered. Retain the droplet rows with explicit curator deference rather than declaring an ACSL3 name-confusion error.
For urinary-exosome PMID:19056867 and B-cell-exosome PMID:20458337, neither full ACSL4 entry was recovered. General abstract descriptions of 1132 and 539 protein surveys do not establish the target-specific detection. Both rows are UNDECIDED. There is also no basis for the old contamination/incidental-cargo assertions.
Human wild-type ACSL4 rescues Drosophila neural/glial and wiring defects in the original abstract of PMID:19617635; patient variants fail rescue. The IBA neuronal-differentiation row remains non-core because it is a developmental consequence in a conserved context, not because the only evidence is disease association. PMID:27656110 provides additional fly lipid/synapse context. PAINT source reviews use ancestral PTN nodes and do not mistake target self-evidence or a short donor list for circularity or weak ancestry.
The insulin-secretion ISS donor O35547 is rat Acsl4. The original rat insulinoma-cell study and PubMed figure captions identify knockdown-dependent changes in stimulated insulin secretion (Fig. 2). These were web-only source checks; no machine cache was created. The cached human UniProt record explicitly assigns this effect by similarity to O35547. Retain the contextual non-core annotation without converting it into a direct secretory mechanism.
PMID:10669417 genuinely resolves to the PLC/PI3K chemoattractant-signaling study in the cached PubMed metadata. Its abstract concerns mouse experiments, but its full ACSL4-linked experiment could not be inspected. The citation-support question remains UNVERIFIED, not MISCITED or WRONG_IDENTIFIER. The original IDA/reference fields are unchanged, and the clearly established human enzyme activity is accepted using independent human experiments.
PMID:12525535 now has a result-bearing P375L/activity quote instead of its title. Primary references retain species and assay scope. The review of ACSL4 (PMID:31306767) is labeled background synthesis. PMID:31504388 is a mouse steroidogenic tissue knockout affecting ester stores, not evidence to create a direct steroid-synthesis core. Recent neuronal, wound and fibroblast disease studies are contextual; the internally difficult direction-of-effect wording in the PMID:39892437 abstract is left unverified and is not used to establish the core regulatory direction.
Provider-only findings and supporting quotes were removed from the curated evidence. In particular, the Cyberian catalytic-triad/residue and blanket localization claims were not accepted simply because a report stated them. Its original file remains immutable and listed as low-relevance, unverified research provenance. Primary observations now support the core and annotation reasons.
Final actions across all 51 original rows: 41 ACCEPT, 4 MODIFY, 3 KEEP_AS_NON_CORE, 3 UNDECIDED. No rows are removed or newly asserted. All 37 references have manual source assessments. The three unresolved rows are the >22-carbon activity and two exosome localizations. The review status remains COMPLETE because all rows have decisions; COMPLETE does not mean every underlying experiment was accessible.
Source-field equality was checked against the original parsed review for every row, excluding only the editable review section. Every primary supporting quote and finding quote was checked as a normalized verbatim substring of its unchanged cache. YAML trailing whitespace was stripped with parsed-YAML equality asserted. Gene validation, history validation and rendering results are recorded in the matching append-only history and the publication manifest.
Final checks passed: just validate human ACSL4, just validate-history history/genes/human/ACSL4/2026-09-26T224730Z-codex-142634.yaml, and just render human ACSL4. The only gene warning recommends citing available deep-research reports; primary evidence was intentionally used instead. There are no validation errors. The four-file publication manifest is /tmp/ACSL4-audit-manifest.json.
The coordinator independently inspected all 51 annotation judgments, the catalytic core, all reference assessments and this source-access record. The biological audit passed. A citation-wide check found that PMID:23766516, cited above for the rat insulin-secretion donor, is not in the local publication cache even though every PMID in the review YAML is cached. A normal ai-gene-review fetch-pmid 23766516 retry on 2026-09-26 returned a DNS resolution failure and cached 0/1 records. No cache was fabricated. This citation remains explicit and the PR must stay draft until its required cache is recovered; gene validation alone does not inspect every citation in notes.
Read the full review comment 5850705850 against published head 41fcbf7e917a74748edb17b5b8bb2a66fd6dd3a9. The three local gene artifacts exactly matched their published byte-based Git blob hashes before editing. All 51 source assertions and their actions are retained: 41 ACCEPT, 4 MODIFY, 3 KEEP_AS_NON_CORE and 3 UNDECIDED.
The palmitoyl-CoA ligase row now attaches the actual cached PMID:9096315 kinetic statement about low palmitate affinity. This paper assayed purified recombinant rat ACS4; the quotation is ortholog evidence supporting secondary substrate use alongside the human Reactome event, not a new claim of purified-human palmitate kinetics. The prior human HUFA-preference quotation did not establish palmitate activity and has been replaced in this row.
The MAM row now rests on the cached full PMID:29450800, Figure 6 and associated MCF-7 Results. The authors isolated MAM, cytoplasmic and mitochondrial fractions and immunoblotted ACSL4 together with calnexin, caveolin, HMGCR and Sigma1R; the caption reports two experiments with similar results. They detected only a minor ACSL4 fraction in MAM, with the larger pool outside MAM and a distribution resembling calnexin. Their conclusion explicitly limits ACSL4's usefulness as a universal MAM marker. The retained contact-site annotation therefore describes a demonstrated partial MCF-7 MAM pool, not exclusive residence or marker specificity in every cell type. PMID:23455425 retains its original source fields and abstract-only cache flag; the previously recorded external HEK293 result remains corroboration. The cached independent MCF-7 experiment now supplies the machine-checkable support. No autophagy-process function follows from this localization.
The unresolved >22-carbon activity row now records the positive scope of PMID:9096315: its rat enzyme panel spans C8–C22 saturated and C14–C22 unsaturated substrates. This panel does not test molecules above C22; that is a limit of the experiment, not a negative result for untested substrates. Independent human DHA/adrenate assays likewise establish C22 activity. Neither panel identifies the substrate of the inaccessible original patient-cell experiment in PMID:11889465. UNDECIDED remains appropriate for that specific very-long-chain activity. The general long-chain ligase row already states its independent human recombinant support, explaining why it can be accepted without resolving the patient assay's exact substrate.
For the biosynthesis suggestion, the cached Reactome R-HSA-75876 describes the actual two-carbon elongation cycle and explicitly includes the activation and condensation steps for arachidonate. R-HSA-548843 identifies ACSL4 as the activation catalyst. A new exact pathway quotation makes this scope explicit: the inference is participation by catalyzing activation within elongation, not an assertion that attaching CoA extends the carbon skeleton. This is a retained existing process assertion, not a NEW annotation.
A bounded comparator check on 2026-09-26 inspected the MGI comparative experimental-annotation tables for Acsl1, Acsl3, Acsl5 and Acsl6. These tables explicitly represent a 2023-03-10 snapshot restricted to experimental annotations. The exact GO:0006633 term appears for rat Acsl3 with IDA; it was not displayed for the other three families. Thus the inspected set does not establish a blanket convention excluding activating ligases. It is not an exhaustive current human all-evidence survey, and missing entries are not evidence of functional absence. The existing local human ELOVL1, HACD2 and TECR reviews each retain source assertions for GO:0030497, GO:0035338 and GO:0006633, consistent with related but distinct steps. Their presence alone does not establish ACSL4's participation; the explicit ACSL4 Reactome reaction supplies that evidence. No comparator genes were edited or newly seeded.
The broad cytoplasm annotation remains ACCEPT at its existing source resolution: the independently demonstrated ER and mitochondrial pools are cytoplasmic structures, and listing narrower core locations does not negate that broader location. The two exosome rows remain UNDECIDED because their ACSL4-specific identifications have not been inspected; a non-core classification would still presume that the target detection is established. The core description now explicitly emphasizes arachidonate alongside other demonstrated C20/C22 substrates. Its broad long-chain ligase term covers that observed substrate range without a redundant separate core entry.
The review status is now DRAFT to match the publication gate already documented above. The notes-only PMID:23766516 cache is still absent; no citation was deleted to hide the gap, and no source cache or provider report was fabricated. Earlier published history remains unchanged. Validation, rendering and source-preservation results are recorded in the follow-up manifest and new history record.
All eight canonical ACSL4 file blobs were checked against main d2d8c9043b082a62378eff620ec0122d4118173b before this follow-up; they matched exactly. The prior merged review and both earlier history records remain unchanged.
The required notes-only source PMID:23766516 has now been recovered by the repository's normal publication fetcher in isolated GitHub Actions run 36286975328, source commit 5946477c8ac79ade0709264c775ea1262b108438. Its unchanged bytes were imported through the verified artifact transport receipt: artifact 10920674630, artifact SHA256 c0ffe4a66b80278af34b44aab6a3ae354ffd5699236b3a486ca95527be5e9713, transport run 36288441414. The publication file SHA256 is 7e6f0b51bb3a051e7861f8b8786aa9f7e61656c1f34a941072d88957c5ea422f; its byte-derived Git blob is 21986a6b4eadf735766e2474bfc4389549c4c357. The cache was not edited or reconstructed by hand.
The normal record is abstract-only (full_text_available: false). Its identity and content agree with the already documented primary source: Acsl4 knockdown in INS 832/13 cells alters EET partitioning and glucose-stimulated insulin secretion; receptor-antagonist treatment partially rescues the response. These observations support the existing contextual regulatory interpretation, not a new claim that ACSL4 executes the secretion machinery. The recovered abstract does not expand prior full-text access or resolve unrelated uncertain annotations.
The notes-inclusive census now has 27 cited PMIDs, all cached. Earlier dated statements that PMID:23766516 was missing are superseded by this recovery record. There are no new biological decisions, reference identities, supporting quotations or core-function changes. The YAML remains DRAFT because the existing unused-provider advisory remains; closing the source-cache publication gate does not by itself satisfy the schema's no-warning COMPLETE criterion. Targeted validation, append-only history validation, rendering and source-byte preservation are checked for the final closure manifest.
The current-head review (formal review 5329551754; detailed comment 5854264401) identified an evidence-attachment gap after the cache recovery. All eight local gene-file blobs were matched to published head 65c401de60a005ffb8af14ac78fdb3d827213bb5 before editing. The normal cached PMID:23766516 abstract was read and is now attached verbatim to GO:0032024, with a new top-level reference assessment and an explicit donor-source citation. The empirical result concerns rat INS 832/13 cells: Acsl4 suppression reduces glucose-stimulated and fatty-acid-potentiated insulin secretion, and the abstract reports altered EET availability and partial rescue by an EET receptor antagonist. It is not a direct human-cell assay. The human ortholog transfer remains KEEP_AS_NON_CORE and does not assign ACSL4 the secretory machinery's function.
The new reference has full_text_unavailable: true, reflecting the actual local abstract-only cache. The existing PMC identifier and the prior external original-paper reading remain distinct from machine cache availability. No source cache is refreshed, rewritten or represented as newly recovered full text. The reviewer suggested checking a possible OA-subset limitation, but neither the existing record nor the evidence inspected establishes why the fetcher obtained only the abstract; no causal explanation is invented. The necessary result is available verbatim in the current abstract.
Published history records remain immutable. The new append-only record records the actual annotation/reference/notes work and links PR #3247 and the preceding review PR #3160. The reference list gains only PMID:23766516; all existing reference objects, 51 annotation source objects and actions, isoform qualifiers and the integrated core remain preserved. All raw UniProt/GOA/provider artifacts and the publication record are unchanged. This bounded follow-up does not reclassify other papers from their PMC identifiers or expand unrelated biological conclusions.
Follow-up validation passed with the pre-existing advisory that no annotation cites the available provider reports. The new abstract quote is an exact case-sensitive match after whitespace normalization; all 51 original source objects and actions, the core and previous reference objects are preserved. The new history and regenerated HTML were checked separately.
All eight canonical gene and provider files matched PR #3247 head cbba19482a5d0f15168c258314430fcbd4be35fb before this session. Its current formal review 5329918321 approves the preceding rat insulin-secretion evidence attachment. This session preserves all 51 annotation source objects, actions, supported quotations, alternative products and the integrated core. No new GO annotation is proposed. The three provider reports and published histories remain byte-identical.
The earlier 27-PMID census covered explicit authored identifiers but omitted DOI-, PMC- and article-URL citations in retained provider reports and older notes. A recursive decoded citation audit found 36 additional real PMID sources. All 36 were fetched normally in source15 run 36310799474, source head a967c4c93e6d01143f35ce7601727d4b9474e9fb. Artifact 10930221362 has SHA256 3ae3ec5148d2565307753a96a3444bda020a2e4cdb9b1acbd8e2b011a4868246. The verified import copied exact generated bytes without overwriting existing records. Its receipt SHA256 is e2300387039c3c58d04c9d9dcff25ab01f57b7b2bf2bc780052b40bddf2ae574; per-file hashes are listed in the publication manifest. Import eligibility and citation identity do not endorse a provider's biological interpretation.
There are 10 abstract-only records and 26 records with actual extracted article or review bodies in this recovered set. Three readers inspected primary identity and available content, and the gene reviewer checked the gene-relevant claims against these records, including original experimental passages for lipid activation, ferroptosis, inhibitor and cancer studies. Body availability does not imply that every supplemental figure, image or raw dataset was inspected. The following source-specific assessments supersede any earlier missing-cache statements for these identifiers.
| Source | Actual local access | Evidence and limits relevant to ACSL4 |
|---|---|---|
| PMID:17110164 | Abstract only | Abstract-only despite PMC. Rat ACSL4 site-directed mutagenesis with fatty-acid selectivity and ATP Km effects; bacterial alignment/templates motivate choices. Rat residue numbering is not a human structure. Provider citation calls this a Yan2015 review, but its PMC identifier maps to the Stinnett2007 primary mutagenesis paper; original provider bytes remain unendorsed. |
| PMID:17259370 | Abstract only | Abstract-only. Human/murine smooth-muscle cells and human macrophages express ACSL4 and show rosiglitazone-sensitive ACSL activity/lipid partitioning; mouse macrophages do not express ACSL4 in this experiment and are not inhibited. Cellular activity and drug correlation are not a purified human ACSL4-specific measurement or proof that every effect is uniquely mediated by ACSL4. |
| PMID:17379924 | Abstract only | Abstract-only. Human HuH7 lipid-droplet fractions show acyl-CoA synthesis and neutral-lipid incorporation; abstract foregrounds ACSL3. It does not settle the ACSL4 supplementary identification, nor contradict curators who read the full paper. Prior external Discussion access is separately recorded in existing notes; this recovered abstract is not full access. |
| PMID:20186809 | Abstract only | Abstract-only human family genetics. Contiguous Xq22.3 deletion spans COL4A5 and ACSL4; chimeric transcript, Western blot and enzyme assay indicate ACSL4 functional nullisomy. Two affected males and multigene deletion are not isolated proof of each clinical feature being caused by ACSL4 alone. |
| PMID:21085606 | Extracted article body | Substantial original body with Methods/Results. MDA-MB-231 and MCF-7 human breast-cancer cell experiments use perturbation and Tet-off ACSL4 expression, AA tracking, LOX-product measurements and COX2 expression. Phenotypic/metabolic regulation is distinct from ACSL4 catalyzing LOX oxygenation, prostaglandin synthesis or acting directly as a transcription factor. Full transgene sequence/species was not independently mapped in this bounded read. |
| PMID:21184843 | Extracted article body | Substantial original Methods/Results. Rat ACSL3/6 variants and rat-liver ACSL4-Flag expressed in E.coli; bacterial lysates with empty-vector controls are used, not purified human protein. VPA inhibits AA/DHA activation by rat ACSL4 with substrate/inhibitor specificity and high millimolar drug concentrations. Clinical efficacy is a hypothesis, not tested human outcome. |
| PMID:22482801 | Abstract only | Abstract-only secondary XLID review despite PMC. Historical counts and discovery narrative are background, not new ACSL4 biochemistry or a current prevalence estimate. No gene-specific detailed evidence extracted from the inaccessible body. |
| PMID:23354024 | Extracted article body | Substantial original Methods/Results. Rat-liver ACSL4-Flag in E.coli lysates with empty-vector controls and radiolabeled AA-CoA readout. PIA inhibition is high millimolar (reported Ki about 11.4 mM) with substrate inhibition and model comparison. This is not purified human enzyme and does not establish safety, anti-bipolar efficacy or in vivo pharmacology. |
| PMID:23520119 | Abstract only | Abstract-only clinical single-patient multigene deletion. A de novo 1.56 Mb deletion includes ACSL4 and eight neighbors but excludes COL4A5. Supports disease context; does not assign every phenotype to ACSL4 or establish a molecular developmental step. |
| PMID:27375556 | Extracted review body | Substantial secondary review body. Summarizes ACTH/kinase/phosphatase signaling, Y1/adrenal and Leydig systems, ACSL4 production of AA-CoA followed by ACOT2 liberation of AA and StAR induction. Prior perturbation studies are cited, not new experiments of this review. ACSL4 is not the thioesterase or steroid-converting enzyme. |
| PMID:27842066 | Extracted article body | Substantial original Methods/Results. Acsl4 knockout is in mouse Pfa1 embryonic fibroblasts; additional mouse lung cells and kidney tissues are distinct from human 15-LOX biochemical assays. Redox lipidomics identifies oxidized arachidonoyl/adrenoyl PE; loss of ACSL4-dependent precursor supply limits ferroptosis. LOX performs oxygenation; this does not establish intrinsic ACSL4 phospholipid acyltransferase or oxygenase activity. PMC title expands PE/AA abbreviations relative to PubMed; PMID/DOI identity remains the same. |
| PMID:33340617 | Abstract only | Abstract-only. HCC cell perturbation/rescue and human-patient correlations connect ACSL4 to cMyc/SREBP1 and lipid/oncogenic phenotypes. In vivo system and exact construct details are not specified by the abstract; no purified enzyme or direct transcription-factor activity inferred. |
| PMID:33634130 | Extracted review body | Secondary MAM/breast-cancer review, including ACSL4 as a marker and prior Leydig-cell work. It supplies no new ACSL4 localization assay; other MAM proteins perform the lipid-transfer, decarboxylase and cholesterol-esterification reactions discussed. |
| PMID:34095144 | Extracted review body | General protein-lipidation review. The ACSL paragraph concerns ACSL1, while DHHC and NMT enzymes perform palmitoylation and myristoylation. ACSL4/FACL4 is not named in the recovered body; this is background rather than ACSL4-specific experimental evidence. |
| PMID:34924561 | Extracted review body | General tumor-microenvironment/post-translational-modification review, without an ACSL4/FACL4/ACSL occurrence in the recovered body. DHHC/PAT and APT enzymes perform the discussed acylation/deacylation. Its linked correction is PMID:35095101; this is not a retraction or an ACSL4 experiment. |
| PMID:35910359 | Extracted review body | Secondary ACSL4 cancer synthesis distinguishes PUFA activation from downstream phospholipid incorporation and describes opposing tumor-survival and ferroptosis effects across contexts. It supplies no new enzymology or clinical intervention and does not assign LPCAT chemistry or lipid oxygenation to ACSL4. |
| PMID:36269835 | Extracted article body | Full Results Figure 1 uses ACSL4 as a marker in adult mouse cardiomyocytes. Genetic, calcium and ischemia experiments primarily test CLIC4; HEK293T cells express CLIC4 constructs and human-heart comparisons concern CLIC4. These are not ACSL4 cardioprotection, chloride-channel or human MAM-topology assays. |
| PMID:36581060 | Abstract only | Abstract-only AA/iron injury study in dopaminergic neuronal models and organotypic slices, with ferrostatin and ACSL4/ALOX inhibitor protection. The abstract does not specify species or establish target-exclusive drug controls; no direct human-neuron or purified-enzyme conclusion is inferred. |
| PMID:36899378 | Extracted article body | Full Methods/Results use human CHON-001 cells, IL-1beta, ACSL4 knockdown and SP1/promoter-reporter perturbations, with no in vivo experiment. Results describe reduced injury and increased viability after ACSL4 knockdown; the title says suppression of ferroptosis, and abstract viability wording also conflicts with Results. These internal directional inconsistencies are not used to define the core. ACSL4 is the regulated target, not SP1 or a DNA-binding transcription factor. |
| PMID:36909239 | Extracted review body | General palmitoylation review without an ACSL4/FACL4 occurrence in the recovered body. DHHC-mediated protein acylation and thioesterase-mediated deacylation are distinct from fatty-acyl-CoA supply. No ACSL4-specific protein-acyltransferase or localization assay is established. |
| PMID:37166352 | Abstract only | Abstract-only LUAD study reports ACSL4/LPCAT3 overexpression and knockout effects on ferroptosis. The YAP/ZEB/EP300 promoter mechanism concerns LPCAT3 transcription. Cell and xenograft context is reported, but detailed construct and host provenance is unavailable; no ACSL4 transcription-factor activity is inferred. |
| PMID:37223023 | Extracted review body | Secondary steroidogenic-signaling synthesis across murine MA-10/Y1, human H295R and rat/bovine systems. ACSL4 supplies arachidonoyl-CoA; Mfn2, OXER1, MKP and steroid-converting enzymes perform their respective functions. The review is not a new human ACSL4 assay, and its loose arachidonate-synthesis and adrenal-zone wording is not used as biochemical evidence. |
| PMID:37372148 | Extracted review body | Secondary ACSL4 cell-death/autophagy review. ACSL4 being an autophagic or proteasomal substrate does not establish that it performs degradation. Specific structural/catalytic-residue claims need original experimental support; this review does not provide a new human structure or purified assay. |
| PMID:37373168 | Extracted review body | Secondary CNS/ferroptosis synthesis explicitly includes rat cells, mouse fibroblasts and EAE mice alongside human observations. This does not support the provider framing of uniformly human evidence. Adrenate is C22:4 despite the review grouping it with C20 substrates; upstream control of ACSL4 is not ACSL4 kinase activity. |
| PMID:37442770 | Extracted review body | Secondary synthesis of ACSL4 acyl-CoA formation, phospholipid remodeling, ferroptosis and disease models. Species and mechanism claims must follow each underlying primary study, not the review heading. |
| PMID:37675456 | Abstract only | Ninety male Sprague-Dawley rats subjected to surgical brain injury; rosiglitazone, ACSL4/transferrin expression, neuronal injury, iron staining, ROS, water content and neurological scoring at 48 h. Supports model-specific pharmacological associations; no human experiment or isolated ACSL4 catalysis is established by this abstract. Do not translate the abstract wording that ROS are produced by ACSL4 into a direct ROS-generating enzymatic function. Drug effects are not proof of target-exclusive causation; unseen full paper not presumed absent. |
| PMID:37884942 | Extracted article body | Female BALB/c bleomycin skin/lung fibrosis model; rosiglitazone intervention, biochemical/histological readouts, mouse bone-marrow-derived macrophages and RAW264.7 with LPS and calpain knockdown/overexpression. Supports a murine calpain/ACSL4-associated ferroptosis/fibrosis model. No direct human ACSL4 experiment located in inspected Methods/Results. Rosiglitazone is explicitly also a PPAR-gamma agonist. Genetic perturbations described here target calpains; pharmacological benefit should not be recast as purified ACSL4-specific catalysis or unqualified human fibrosis function. |
| PMID:38471082 | Extracted article body | Human TNBC lines including MDA-MB-231, BT-549, Hs578T; human breast tumor samples; human tumor cells in NCG mice. ACSL4 knockdown/re-expression and lipidomics support membrane phospholipid remodeling; downstream CD47-dependent integrin beta1 raft localization/activation and FAK signaling accompany migration/metastasis. Do not convert the downstream integrin/FAK effects into direct ACSL4 receptor binding or kinase catalysis. Mouse host and human tumor-cell identity are distinct. |
| PMID:38552012 | Extracted article body | Human HT-1080/MDA-MB-231/HEK293T experiments and explicitly human O60488-1 ACSL4 re-expression; additional human/mouse cell lines; C57BL/6 male mouse kidney I/R and acute liver injury. Recombinant ACSL4 binding to biotinylated AS-252424 and inhibition of AA-to-AA-CoA activity are measured. Q464 site inference combines AlphaFold docking, mutant drug-response/binding tests; it is not an experimentally solved ligand-bound structure. Recombinant protein production species/host is not explicit in the inspected activity-method paragraph, although human isoform 1 is explicit for rescue. Do not silently promote the whole assay panel to one construct/host. Source identity/body are coherent; pharmacological specificity and model efficacy are reported findings, not clinical efficacy proof. |
| PMID:39247444 | Extracted article body | Human patient transcript datasets plus human Caco-2 intestinal epithelial cells. ACSL4 knockdown or rosiglitazone changes oxidative/antioxidant markers, cell viability/apoptosis and NF-kappaB inflammatory readouts in LPS stress. The Methods DEG threshold states /log2FC/ >2 whereas Results/Fig1 use >1 and ACSL4 log2FC=1.178; this is a source-internal analysis-description inconsistency. The explicitly measured Annexin V/PI endpoint is called apoptosis, and oxidative markers alone do not uniquely diagnose ferroptosis. Retain the authors reported model interpretation with assay limitations; no purified ACSL4 chemistry or in vivo human treatment is shown. |
| PMID:39496916 | Extracted review body | Secondary synthesis across ferroptosis/stroke models and prospective ACSL4 interventions; no new gene-specific experimental Methods/Results. |
| PMID:40766753 | Extracted review body | Broad cerebral ischemia/reperfusion therapy review; ACSL4 occurs in cited miRNA/exosome/ferroptosis pathways. No new ACSL4 construct, activity or animal experiment by these authors. |
| PMID:40932861 | Extracted review body | Secondary ACSL-family/lipid-peroxidation overview. Organelle and cancer claims aggregate underlying sources; no newly measured ACSL4 localization or enzymology. The body states ACSL4 synthesizes PUFA-CoA by adding acetyl-CoA to free fatty acids; this is incorrect reaction wording. ACSL4 ligates free fatty acid to CoA using ATP, not acetyl-CoA. Preserve source bytes but do not propagate this wording. Provider journal-quality labels are not independently endorsed. |
| PMID:41288931 | Extracted review body | Secondary review of five ACSL isoforms and clinical/preclinical studies. Family-wide summaries are not automatically ACSL4-specific primary evidence. |
| PMID:41383617 | Extracted review body | Secondary Parkinson disease/ferroptosis synthesis across animal and cell models and clinical developments; no new primary ACSL4 assay. |
| PMID:41425599 | Extracted review body | Secondary neurodegeneration/ferroptosis review, including cited genetic, biomarker and pharmacological model studies; no new ACSL4 experimental Methods/Results. |
The additional biochemical work corroborates fatty-acid activation and lipid remodeling. Rat constructs in bacterial lysates are not purified human enzyme; human tumor cells implanted in mice are not mouse ACSL4 constructs; pharmacological benefit is not automatically target-exclusive causation. The mouse lipidomics study PMID:27842066 assigns phospholipid oxygenation to lipoxygenases and precursor supply to ACSL4. The human TNBC study PMID:38471082 links membrane lipid remodeling to downstream integrin/FAK signaling without making ACSL4 a receptor or kinase. No new intrinsic protein-palmitoylation, oxygenase, autophagy-machinery, membrane-fusion or steroid-converting function follows from the recovered references. No new evidence resolves the three existing UNDECIDED rows.
PMID:17379924 is the real bibliographic identity of the older JLR article URL in these notes. Its recovered abstract emphasizes ACSL3 in HuH7 droplet fractions and measures bulk ACSL activity; it does not newly recover the ACSL4-specific full-paper observation. The previously documented external Discussion read and curator deference remain separate. The source cache is not represented as full text, and the existing droplet judgments are unchanged.
The unchanged reports contain three wrong identifiers: PMID:9598322 is a honeybee library paper rather than the intended human-cloning PMID:9598324; PMID:11889466 is an ATM-splicing paper rather than FACL4 PMID:11889465; PMID:33359517 is a glioma-printing paper rather than ACSL4/c-Myc PMID:33340617. These are documented exceptions, not additional ACSL4 cache requirements. The reported DOI prefix 10.1093/cm9.0000000000002533 is also wrong; the primary PMID:37442770 record gives 10.1097/CM9.0000000000002533. No source file was edited to hide these mistakes. General PTM references PMID:34095144, PMID:34924561 and PMID:36909239 do not establish ACSL4-specific protein-acyltransferase activity. PMID:40932861 has incorrect acetyl-CoA reaction wording; the ligase uses fatty acid, CoA and ATP. PMID:36899378 has internal title/abstract/Results direction inconsistencies that are explicitly recorded rather than copied into the biological summary.
The original Wiley open-access review by Zhang and Wang, DOI 10.1002/ctd2.200, was read directly on 2026-09-27, including its ACSL4 section. It summarizes ACSL-family metabolism and cancer studies and presents no new ACSL4 experiment. Its broad organelle and signaling summaries are secondary evidence; the underlying species and enzymes must remain distinct. The separately indexed original Neurology meeting-supplement abstract, DOI 10.1212/WNL.92.15_supplement.P5.6-038, was also read. It reports a five-year-old boy with a hemizygous frameshift and carrier mother/sister, without an enzyme assay. It provides clinical context rather than a measured molecular function. No PMID was established for either work, and neither is a fabricated PMID cache or a preprint. Their verified publisher access is recorded separately from the 36 recovered PubMed records.
The official PubMed record for PMID:35095101 identifies the correction to PMID:34924561, DOI 10.1038/s41392-022-00901-7. The original publisher correction was read through its indexed full text on 2026-09-27; a direct PDF request at this follow-up encountered the publisher redirect. It corrects chemical structures in Figure 1, an author name and protein-glycosylation wording, with the authors stating that the key findings are unaffected. It does not retract the article or add an ACSL4 assay. One normal just fetch-pmid 35095101 attempt completed with exit 1, DNS failure and 0/1 records; no cache was fabricated.
The final required census therefore has 64 distinct valid PMID citations: 63 cached and one missing correction, plus four cached Reactome records and the two separately verified DOI-only works. The bibliographies of those papers are not recursively promoted into new gene-review citations. Unasserted raw UniProt bibliography entries and the three wrong provider IDs are distinguished from genuine authored/provider requirements. The review and PR remain DRAFT for PMID:35095101; the pre-existing unused-provider advisory is also retained. Closing 36 source gaps is not a blanket draft-to-ready claim. Targeted validation, new-history validation, rendering, exact source/quote preservation and per-file hashes are recorded in the session manifest.
The sole missing PMID record, PMID:35095101, is now present as exact normal-fetch bytes from source18. It contains bibliographic metadata only, without an abstract or correction body; the local full_text_unavailable: true flag remains accurate. This supersedes the earlier missing-cache statement without expanding the cache's content.
The indexed publisher correction was independently read again. Besides Figure 1 chemical structures and the author-name spelling, it corrects the O-GlcNAcylation section's glycosylation definition on page 6, lines 45–47, to proteins rather than proteins and lipids, and removes the eukaryote restriction. The authors state that key findings are unaffected. Individual corrected structure drawings were not visually inspected. No ACSL4 assay or annotation change follows from this notice.
Current-head review 5330408790 and comment 5856073399 were read in full. The general PTM review PMID:34924561, general protein-palmitoylation review PMID:36909239, and linked correction are retained with relevance NONE to document the provider bibliography, rather than as ACSL4 evidence. The choice reflects their actual use here; LOW can still cover useful general background without requiring a literal gene-name occurrence. The source-internal reaction error in PMID:40932861 and directional/reporting inconsistencies in PMID:36899378 now carry LOW_QUALITY, with the specific defect bounded in each reference assessment. Neither a miscitation by this review nor an independently established external dispute is asserted.
All 51 annotation decisions, core functions, source assertions, alternative products and 82 supporting quotations remain unchanged. The suggested additional quotes from PMID:38552012 and PMID:38471082 are not required to repair existing support: their actual assay scope is already documented in the source15 assessments, and the existing ligase and ferroptosis evidence remains intact. Reference-assessment rendering is a shared template issue outside this gene follow-up. Five spacing errors in authored prose were corrected. The previous citation-format normalization preserved scientific text and source destinations while avoiding duplicated links in rendered notes; this paragraph records that earlier formatting-only journal change explicitly.
The recursive authored/provider census remains 64 valid PMIDs and four Reactome records, now all cached, plus the two previously assessed DOI-only works with no established PMID. The three identified wrong provider PMID citations remain explicitly excluded rather than silently remapped. No new source was requested. Citation closure and schema status are distinct: DRAFT is retained if the pre-existing unused-provider advisory remains after validation; it does not indicate a missing publication cache.