This guide separates inspected source evidence from remaining access limits. Normal publication-cache flags are preserved; several originals were recovered separately while their normal cache remains abstract-only. Reference-level full_text_unavailable flags are omitted where operative original text was recovered; partial reads, missing figures and missing supplements remain specified below. This does not change the independent normal-cache full-text flags. Original article bodies, figure images and private workspace paths are not redistributed here. Numerical row indices below are zero-based positions in the frozen 104-annotation review.
Publisher DOI; accessed author copy. Original cellular affinity-labeling, crosslinking, immunoprecipitation and competition Methods/Results, Discussion and captions; saved original lines 577–1557. Later citing-paper text on the same page is excluded.
Independently assessed original affinity-labeling, crosslinking, immunoprecipitation and competition experiments support TGFB1/TGFB3 association in human HUVEC and human-ENG COS-M6 receptor systems. TGFB3 was tested as competitor. This is not a purified binary-affinity assay. A separate frozen UniProt prose inversion does not invalidate the correct imported ligand partners.
Reviewed annotation contexts: 27, 43, 44, 45, 46, 47, 48, 49, 53, 54, 83, 84, 85. Normal cache full-text flag: false.
Access locator: authenticated PubMed abstract and normal metadata only. No operative original-paper Methods, figure or supplement assessment is claimed.
Regulated expression on human macrophages of endoglin, an Arg-Gly-Asp-containing surface antigen.
Authentic abstract explicitly reports endoglin-specific flow cytometry and surface expression of differentiated human monocytes/macrophages. Full operative Methods were unavailable. This supports the external membrane context, independently corroborated by topology and surface-labeling originals; no RGD-dependent adhesion function is inferred.
Reviewed annotation contexts: 53. Normal cache full-text flag: false.
Access locator: authenticated PubMed abstract and normal metadata only. No operative original-paper Methods, figure or supplement assessment is claimed.
Primary structure of endoglin, an RGD-containing glycoprotein of human endothelial cells.
Authenticated abstract connects human endothelial cDNA with placental endoglin N-terminal sequence and extracellular, membrane-spanning and cytoplasmic architecture. Historical numbering is not forced onto the current precursor. Full operative Methods and exact insert sequences were not examined; the RGD motif alone is not functional adhesion evidence.
Reviewed annotation contexts: 45, 49, 53. Normal cache full-text flag: false.
Publisher DOI; accessed author copy. Operative Methods, Results and Discussion on lectin affinity and carbohydrate treatment; Figure 1 caption. Saved indexed original: Methods lines 624–1620; Results 1622–1853, 2174–2746 and 2800–3357; Figure 1 caption 1970–2031; Discussion 3399–4121.
Biochemical characterization of the 44G4 antigen from the HOON pre-B leukemic cell line.
The genuine original operative Methods, Results and Discussion characterize glycosylated human 44G4/ENG antigen through lectin affinity, galactose competition, glycosidases and antibody recognition. The lectin recognizes the sugar. These experiments do not support assigning galactose/glycosaminoglycan recognition to ENG itself. Positive glycosylation findings are retained; no universal sugar-nonbinding claim or figure-pixel assessment.
Reviewed annotation contexts: 38, 39. Normal cache full-text flag: false.
Access locator: authenticated PubMed abstract and normal metadata only. No operative original-paper Methods, figure or supplement assessment is claimed.
Authentic abstract establishes human ENG-related HHT, but the operative original was not recovered. The specific general and CNS de novo-vasculogenesis assays remain unresolved. Clinical necessity is not substituted for execution of a vessel-formation step, and no wrong-gene or curator-error claim follows from unavailable full text.
Reviewed annotation contexts: 2, 4, 60. Normal cache full-text flag: false.
Access locator: authenticated PubMed abstract and normal metadata only. No operative original-paper Methods, figure or supplement assessment is claimed.
Authentic abstract identifies murine ENG cloning, disulfide-linked homodimer and TGF-beta binding. It is not presented as the later knockout developmental assays; those are explicitly supplied by independently assessed mouse originals. Exact source-specific developmental claims and complete Methods were not directly verified here.
Reviewed annotation contexts: 1, 7, 71, 78, 80, 82, 83. Normal cache full-text flag: false.
Publisher DOI; accessed author copy. Selected Methods on porcine endothelial receptors, radioligand crosslinking and immunoprecipitation; receptor-association Results and Discussion. Saved indexed original: Methods lines 786–944 and 1031–1311; Results 1726–1951 and 2283–2713; Discussion 2770–2808, 2854–2898 and 3166–3245.
Selected genuine original text supports porcine endothelial receptor/endoglin complexes, ligand crosslinking, dimerization and phosphorylation. The specific glycosaminoglycan-recognition assay behind that donor assertion remains unresolved. The historical 643-residue clone is not equated to current 653-residue P37176 or human products. Receptor findings remain positive; no figure pixels or supplement assessment.
Reviewed annotation contexts: 23, 40, 66. Normal cache full-text flag: false.
Access locator: authenticated PubMed abstract and normal metadata only. No operative original-paper Methods, figure or supplement assessment is claimed.
Authentic abstract reports separate human long/short cDNAs expressed in mouse fibroblasts with TGFB1 binding and disulfide-linked dimers. The operative original and exact inserts remain inaccessible. Neither equivalent native tail functions nor quantitative binding equivalence is inferred; current natural products come from the official sequence record.
Reviewed annotation contexts: 47. Normal cache full-text flag: false.
Publisher DOI; accessed author copy. Receptor-expression/affinity-labeling Methods; ligand/receptor Results; Figures 5, 6 and 9 captions and matching Discussion. Saved indexed original: Methods lines 512–741; Results 1977–2012 and 2140–2278; captions 2052–2131, 2280–2349 and 3151–3188; Discussion 2652–2769.
Selected genuine original receptor-expression, affinity-labeling, crosslinking, immunoprecipitation and Discussion text establish ligand association through appropriate ligand-binding receptors. The three imported contributes_to qualifiers are respected. Human constructs and COS-1/NCTC hosts remain distinct; no purified autonomous affinity, fixed native stoichiometry or full insert alignment is claimed.
Reviewed annotation contexts: 34, 35, 36, 37, 69, 72, 73, 76, 83, 86, 87. Normal cache full-text flag: false.
Publisher DOI; accessed author copy. Original knockout design, developmental Results, Discussion and Methods notes: saved original lines 195–439 and 484–709; captions 501–615 and 754–792; Methods notes 800–866. Corrected line indexes preserve all original captured responses; the unrelated article beginning at line 901 is excluded.
Defective angiogenesis in mice lacking endoglin.
Separately recovered original knockout Results, Discussion, captions and Methods notes were assessed. Early endothelial differentiation and primitive vascular networks persist; later remodeling, branching and smooth-muscle investment fail. This distinguishes developmental stages without claiming universal absence of ENG participation. Text only; the adjacent unrelated article is excluded.
Reviewed annotation contexts: 0, 1, 2, 3, 5, 6, 7, 77, 78, 79, 80, 81, 82. Normal cache full-text flag: false.
Access locator: authenticated PubMed abstract and normal metadata only. No operative original-paper Methods, figure or supplement assessment is claimed.
Authentic abstract explicitly describes patient HUVEC/monocyte protein analysis and transfectant surface biotinylation, separating normal surface protein from intracellular missense mutants. The full operative original was unavailable. Supports the bounded localization claim; wild-type and mutant trafficking are not conflated.
Reviewed annotation contexts: 43. Normal cache full-text flag: false.
Original paper. Methods on targeted Eng disruption and embryos; Results on embryonic vasculature and heart; corresponding Discussion and Figures 1–4 captions. Saved original blocks: Methods 22–28; Results 29–39; Discussion 46–53. Figure pixels were not examined.
A murine model of hereditary hemorrhagic telangiectasia.
The genuine JCI original Methods, Results, Discussion and captions were assessed separately from its abstract-only normal cache. Mouse Eng-null embryos initially form primitive vessels, followed by branching/remodeling failure, vascular fragility and a dilated umbilical vein. No direct human construct or universal absence-of-function inference; no figure pixels inspected.
Reviewed annotation contexts: 2, 3, 81, 82. Normal cache full-text flag: false.
Access locator: authenticated PubMed abstract and normal metadata only. No operative original-paper Methods, figure or supplement assessment is claimed.
Identifier and accessory-receptor review abstract are authentic, but the operative review body was not recovered. Its BMP and activin statements are retained with explicit independent primary support; the abstract is not presented as the underlying ligand-binding experiment.
Reviewed annotation contexts: 62, 76. Normal cache full-text flag: false.
Publisher DOI; accessed author copy. Domain-construct and reporter Methods; receptor-association and Smad reporter Results; Figures 3, 4, 5, 9 and 10 captions. Saved indexed original: Methods lines 452–464, 529–607, 1051–1092 and 1133–1175; Results 1177–1262, 1604–1621, 1826–1972, 2163–2222 and 3168–3230; Discussion 3313–3420 and 3944–4086.
Selected genuine original Methods, receptor-domain association Results, reporter Results, Discussion and captions were assessed. Human ENG constructs associate with type I/II receptors and influence ligand-driven reporter activity in distinct COS-7 and rat L6E9 hosts. Receptors phosphorylate ENG, not the reverse. Engineered domains do not establish natural-product equivalence; no figure pixels or separate supplement read.
Reviewed annotation contexts: 21, 22, 23, 24, 25, 26, 51, 52, 54, 58, 59, 64, 65, 66, 91. Normal cache full-text flag: false.
Access locator: authenticated PubMed abstract and normal metadata only. No operative original-paper Methods, figure or supplement assessment is claimed.
Authentic abstract and metadata identify the mouse Eng-null arteriovenous-boundary study. Original Methods, the negative cushion-transition assay, precise repression target and dorsal-aorta assay remain unavailable. Those source-specific assertions remain UNDECIDED; access failure does not establish a curator error or biological contradiction.
Reviewed annotation contexts: 55, 56, 67, 68, 93. Normal cache full-text flag: false.
Access locator: authenticated PubMed abstract and normal metadata only. No operative original-paper Methods, figure or supplement assessment is claimed.
The authentic abstract reports ENG/ALK1/ALK5 association and functional interplay, but operative full text was not recovered. Independent human receptor-domain and signaling originals support the retained broad roles and interaction classes. Exact source-specific assay geometry and every partner experiment remain unverified.
Reviewed annotation contexts: 24, 25, 26, 51, 59. Normal cache full-text flag: false.
Publisher DOI; accessed author copy. Expression-plasmid and BRE reporter Methods; BMP9/ALK1/ENG Results and Figure 4 caption. Saved indexed original: Methods lines 477–625, with human ENG plasmid 541–560 and reporter assay 562–588; Results 1025–1059; Figure 4 caption 1149–1166.
Selected original Methods, Results and Figure 4 caption identify human ENG expression plasmids and increased low-dose BMP9/ALK1 BRE reporter activity in mouse NIH-3T3 hosts. The BMP2 comparison was not significant. Human construct and host species are kept distinct; exact natural-product assignment and figure pixels were not assessed.
Reviewed annotation contexts: 62, 63. Normal cache full-text flag: false.
Access locator: authenticated PubMed abstract and normal metadata only. No operative original-paper Methods, figure or supplement assessment is claimed.
Authentic abstract reports endoglin association with beta-arrestin2 and internalization. The full operative original was unavailable. Independent surface-labeling evidence supports the imported membrane location; internalization is not generalized to constitutive vesicle residence or a new natural product.
Reviewed annotation contexts: 46. Normal cache full-text flag: false.
Original paper. Methods on the human transgene, neural-crest lineage and mouse/rat cultures; Results on cardiovascular phenotypes and myogenic differentiation; Discussion and embedded captions. Saved original blocks: Methods 21–39; Results 40–76; Discussion 77–88. Negative TUNEL/phosphohistone-H3 findings are described as data not shown.
Endoglin is required for myogenic differentiation potential of neural crest stem cells.
Genuine original Methods, Results, Discussion and captions were assessed. Human ENG overexpression in mouse neural crest, mouse-null cultures and rat NCSC assays remain distinct. Smooth-muscle differentiation and compound ENG-overexpression/Smad4-loss outflow phenotypes are supported. Negative TUNEL, phosphohistone-H3 and lineage-distribution findings retain their bounded contexts; no figure pixels or universal negatives claimed.
Reviewed annotation contexts: 9, 10, 57, 77, 78, 89, 90, 94, 95, 98, 99, 102, 103. Normal cache full-text flag: true.
Access locator: authenticated PubMed abstract and normal metadata only. No operative original-paper Methods, figure or supplement assessment is claimed.
Authentic abstract identifies long/short endoglin as auxiliary receptors and reports endothelial senescence, receptor interactions and differing signaling responses. Operative Methods, exact constructs and natural-tail equivalence remain unverified. Independent receptor originals support refinement to coreceptor activity.
Reviewed annotation contexts: 21. Normal cache full-text flag: false.
Original paper. Selected operative main-text Methods/Results relevant to the finding above. The retained extracted-text block positions are 21, 24, 25, 26, 29, 33, 34, 35, 38, 39, 42, 44, 47, 50; these positions are a saved-extraction aid, not publication page numbers. No figure pixels or separate supplementary files were examined.
Endoglin (CD105) as a urinary and serum marker of prostate cancer.
Selected authentic main-text Methods and Results report ELISA detection in human post-DRE urine supernatant and serum. Supports an extracellular pool without identifying a unique cleavage species, tissue of origin or extracellular activity. No new tumor process, figure-pixel or supplement claim.
Reviewed annotation contexts: 41, 42. Normal cache full-text flag: true.
Publisher DOI; accessed author copy. Methods on chimeras, embryoid bodies, AVC explants and expression; corresponding Results and Discussion, including figure captions. Saved original lines: Methods 342–623; Results 625–1504; Discussion 1506–1815. Supplementary files and figure pixels were not inspected.
Original mouse chimera, embryoid-body and AVC-explant Methods/Results/Discussion were independently assessed. Eng-null cells can enter early sprouts, but contribute poorly to cushion mesenchyme; delayed transformation retains residual activity. Chamber and expression phenotypes are contextual, and trabeculation may include secondary hypoxia. Corrected saved line locators preserve original responses. No figure pixels or supplements assessed.
Reviewed annotation contexts: 5, 6, 7, 8, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 74, 75, 96, 97, 100, 101. Normal cache full-text flag: false.
Original paper. Selected operative main-text Methods/Results relevant to the finding above. The retained extracted-text block positions are 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64; these positions are a saved-extraction aid, not publication page numbers. No figure pixels or separate supplementary files were examined.
Endoglin phosphorylation by ALK2 contributes to the regulation of prostate cancer cell migration.
Selected authentic main text compares human long ENG and cytoplasmic mutants in prostate-cancer-cell migration assays. Receptor phosphorylation and ligand context modify the inhibitory migration effect. This is not a universal antimigratory function, a short-product assay, or ENG kinase activity; no figure pixels or separate supplements assessed.
Reviewed annotation contexts: 61. Normal cache full-text flag: true.
Publisher DOI; accessed author copy. Selected ENG-perturbation/signaling Results, Discussion and matching cell/affinity-labeling Methods. Saved indexed original: Results lines 482–600 and 634–711; Discussion 1145–1410; Methods 1412–1503. The relevant supplementary affinity-labeling panel was not recovered.
Critical role of endoglin in tumor cell plasticity of Ewing sarcoma and melanoma.
Selected genuine original Results, Discussion and Methods assess ENG perturbation and signaling in human Ewing-sarcoma/melanoma cells. The supplementary affinity-labeling panel needed for full source-specific TGFB1 detail was not recovered. Independent human ligand evidence supports the imported association; TGFB3 experiments do not rewrite its TGFB1 partner.
Reviewed annotation contexts: 27. Normal cache full-text flag: false.
Original paper. Microparticle Results and Figure 1E caption, with corresponding injury, blood fractionation and CD105-imaging Methods. Saved original blocks: Results 33–35; Figure 1 caption 34; Methods 58–63. The separate CD36 signaling experiments are not represented as ENG perturbations.
CD36 participates in a signaling pathway that regulates ROS formation in murine VSMCs.
Focused original JCI Results, Figure 1 caption and matching injury/fractionation Methods show endogenous CD105 in mouse blood-derived microparticle pellets using staining and dot immunoblot. The CD36-focused title does not exclude the target assay. Supports fraction localization, not human isoform specificity, exosome identity or ENG-dependent microparticle production; no figure pixels assessed.
Reviewed annotation contexts: 92. Normal cache full-text flag: false.
Original paper. Focal-adhesion-enrichment Methods and proteomic Results; genuine supplemental workbook Table 1 Excel row 210, Table 2 row 160, Table 4 row 135 and Table 7 row 195. Table 9 protein-group zero-based rows 4668, 7841, 7471, 4722, 4908 and 5671 include matching peptide groups. Validation sheets 5/6 supplied no ENG target row.
Original focal-adhesion preparation Methods and genuine workbook target cells identify ENG/endoglin under legacy IPI00017567.3 in reproducible HFF1 substrate-bound fractions. The long-precursor name is not product-specific peptide proof. Stress-fiber fragments and contamination limits remain explicit; no target validation in the inspected validation sheets or target-specific imaging is claimed.
Reviewed annotation contexts: 50. Normal cache full-text flag: true.
Original paper. Operative Experimental Procedures, Results and Discussion, all main figure captions and main Table 1; saved original blocks 9–55. Separate supplementary files and figure pixels were not read.
Independently assessed original Methods, binding Results and literal affinity tables support direct human soluble ENG recognition of BMP9 and BMP10. Fc/His constructs, assay geometry and ligand-dependent affinities remain distinct. Soluble sequestration can inhibit signaling and does not invert the membrane coreceptor role; no full-length natural-tail equivalence.
Reviewed annotation contexts: 31, 41. Normal cache full-text flag: true.
Original paper. Operative Results, Discussion, Materials and Methods, all main figure/table captions, both main table grids and four supplementary captions embedded in the JATS; saved original blocks 2–45. The genuine recovered JATS route was https://www.ebi.ac.uk/europepmc/webservices/rest/PMC3275592/fullTextXML. Standalone supplementary data and figure pixels were not inspected.
Structural and functional insights into endoglin ligand recognition and binding.
Independently assessed original structural/domain and binding evidence supports human BMP9 recognition with soluble engineered constructs. The earlier truncation-based Cys350 interpretation is kept separate from the later C350-C382 intramolecular bond structure. No complete full-length membrane receptor complex or natural-tail equivalence is claimed.
Reviewed annotation contexts: 31, 41. Normal cache full-text flag: true.
Original paper. Selected operative main-text Methods/Results relevant to the finding above. The retained extracted-text block positions are 20, 24, 25, 26, 28, 30, 32, 34, 35, 37, 39, 41, 45, 47, 48, 50, 52, 54, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77; these positions are a saved-extraction aid, not publication page numbers. No figure pixels or separate supplementary files were examined.
Endoglin mediates fibronectin/α5β1 integrin and TGF-β pathway crosstalk in endothelial cells.
Selected genuine Methods/Results support reciprocal transfected and endogenous ENG-alpha5/beta1 integrin association, proximity and internalization. Human endothelial, mouse endothelial and COS-7 systems are distinguished. Association alone is not an autonomous ENG adhesion function or isolated binary interface; no figure pixels or separate supplement read.
Reviewed annotation contexts: 28, 29. Normal cache full-text flag: true.
Original paper. Selected endothelial-cell, phospho-SMAD, siRNA, antibody and BLI Methods/Results; Figures 1–3 signaling comparisons and Figure 5 antibody/BMP9 binding, with tube-formation/growth context. Supplemental captions were read, not separate trace files or pixels.
Independently assessed operative original evidence links human endothelial ENG perturbation to BMP9 signaling. Receptor/antibody and cellular contexts are retained. This supplies membrane-signaling support alongside direct ligand-binding studies without assigning receptor kinase activity to ENG or equating all natural products.
Reviewed annotation contexts: 21, 58, 59, 62, 91. Normal cache full-text flag: true.
Original paper. Selected operative main-text Methods/Results relevant to the finding above. The retained extracted-text block positions are 30, 31, 32, 33, 34, 37, 61, 71, 83, 89, 90, 93, 96; these positions are a saved-extraction aid, not publication page numbers. No figure pixels or separate supplementary files were examined.
LRG1 promotes angiogenesis by modulating endothelial TGF-β signalling.
Selected genuine original text supports LRG1 association through endothelial coimmunoprecipitation, soluble-domain assays and SPR. A Methods insert-name inconsistency is recorded rather than claiming exact sequence verification. Positive association is retained without inventing an additional ENG catalytic activity; no figure pixels or separate supplement read.
Reviewed annotation contexts: 30. Normal cache full-text flag: true.
Original paper. Selected operative main-text Methods/Results relevant to the finding above. The retained extracted-text block positions are 24, 25, 46, 57, 71, 72, 73, 90, 91, 94, 95, 96, 102, 103, 104, 105; these positions are a saved-extraction aid, not publication page numbers. No figure pixels or separate supplementary files were examined.
Selected authentic original Methods/Results show human ENG 3-prime-UTR targeting by miR-342-5p and reduced HUVEC Akt phosphorylation with rescue. Methods specify murine Eng rescue cDNA. This is not isolated human-protein rescue or ENG kinase activity, and all miRNA effects are not attributed solely to ENG.
Reviewed annotation contexts: 88. Normal cache full-text flag: true.
Original paper. Selected construct, crystallography, mutagenesis, binding and SEC-MALS Methods/Results; Figure 2 OR–BMP9 recognition, Figure 4 dimerization/stoichiometry, Figure 5 ternary-complex experiments and Table 1 BLI values in extracted text. Full ectodomain/membrane assembly is modeled, not a solved complete receptor complex; no independent table-grid, supplement or pixel assessment.
Structural Basis of the Human Endoglin-BMP9 Interaction: Insights into BMP Signaling and HHT1.
Independently assessed original structural, binding and mutational evidence resolves human orphan-region BMP9 recognition and distinguishes ZP-mediated covalent dimerization. Folding/secretion controls and C350-C382 intramolecular bonding are considered. Soluble complexes are not a complete membrane-receptor assembly, and proposed type II receptor handoff is not an observed step. No natural-tail equivalence or figure-pixel claim.
Reviewed annotation contexts: 21, 25, 31, 58, 62, 70, 71. Normal cache full-text flag: true.
Original paper. Selected operative main-text Methods/Results relevant to the finding above. The retained extracted-text block positions are 20, 23, 26, 29, 30, 31, 34, 37, 40, 43, 50, 51, 54, 55, 56, 57, 60, 61, 62, 63, 64, 67, 68, 69, 72; these positions are a saved-extraction aid, not publication page numbers. No figure pixels or separate supplementary files were examined.
Endoglin Protein Interactome Profiling Identifies TRIM21 and Galectin-3 as New Binding Partners.
Selected authentic main text supports galectin-3 and TRIM21 interactions by array screening, coimmunoprecipitation and recombinant-protein BLI. Galectin-3 recognizes glycan-bearing ENG; ENG is not thereby a lectin. Reciprocal endogenous TRIM21 evidence is not uniformly positive, and its proposed ENG degradation mechanism remains future work.
Reviewed annotation contexts: 32, 33. Normal cache full-text flag: true.
The two official natural products are Long P17813-1 (658 residues) and Short P17813-2 (625 residues). The VSP_004233 replacement changes long-product residues 619–658 to EYPRPPQ. Signal peptide 1–25, mature-chain processing and experimental soluble fragments are separate concepts. Both products share the extracellular sequence; this is not evidence that their distinct cytoplasmic tails are functionally interchangeable.
The 106 raw QuickGO records yield 105 normal TSV rows and 104 seeded annotation objects. Row 14 retains two original date/extension contexts; row 75 retains two tissue/input contexts that collapse in the normal projection. The archive preserves the original records. Five NOT annotations and three contributes_to annotations remain unchanged.
The imported ligand partners in rows 84/85 are correct TGFB1/TGFB3 contexts. A separate frozen UniProt SUBUNIT sentence reverses TGFB2 and TGFB3 relative to the inspected 1992 original. This is a record-text question; it is not a defect attached to those correct annotations and no upstream report has been sent.
The glycan judgments concern the direction of recognition in PMID:3262645. They retain positive ENG glycosylation and lectin-association findings and do not assert universal inability of ENG to bind sugars. The porcine glycosaminoglycan assertion remains unresolved.
The corresponding reproducible analysis reconstructs the machine-source context. It does not automate scientific annotation decisions or replace original-article assessment.