COL18A1

Summary: The skeletal-development IBA has positive descendant evidence, but its target-specific scope is unresolved.

Reason: The ZFIN source resolves to col27a1a; original PMID:20041163 reports controlled zebrafish knockdown experiments affecting notochord integrity and vertebral development. The mouse source resolves to Col27a1. These are genuine positive collagen-family observations, not evidence that the donor annotation is wrong. The historical PTN008601313 placement and a COL18A1-performed skeletal-development step were not resolved in the inspected sources. Neither donor count nor the different collagen subtype establishes an erroneous propagation; retain the original assertion as undecided rather than infer a phylogenetic failure. PMID:10942434 provides positive human Knobloch-syndrome evidence, including occipital encephalocele and a neural-tube-closure interpretation. Its available abstract does not distinguish a skeletal-development contribution from the adjacent neural/cranial developmental phenotype. That unresolved specificity does not negate structural participation or establish that the historical IBA node was misplaced.

Summary: Collagen XVIII and its processed forms influence angiogenesis in specific developmental and experimental settings.

Reason: Mouse Col18a1 has experimental vascular-development evidence, including altered retinal vascular regression in PMID:11927538 and positive effects of added endostatin in embryonic-body assays in PMID:15188432. Human oligomeric NC1 has matrix-dependent motogenic effects in PMID:11257123, while PMID:9008168 reports inhibitory endostatin preparations. These support the broad inherited process without assigning one regulatory sign or equivalent activity to every precursor and fragment. Vascular responses remain contextual to the structural core.

Summary: Endothelial morphogenesis is supported for particular NC1/endostatin preparations.

Reason: The mouse donor carries a direct morphogenesis annotation from PMID:15188432. Original human NC1 and engineered endostatin dimers in PMID:11257123 alter endothelial movement and morphogenesis on Matrigel; monomeric endostatin does not reproduce the motogenic effect and can antagonize oligomers. This is positive performed extracellular activity, but not a universal intact-precursor activity or a separately established receptor/kinase mechanism.

Summary: Human collagen-XVIII-derived fragments occur extracellularly in plasma.

Reason: PMID:10441114 explicitly identifies glycosylated human plasma endostatin and circulating related fragments. Extracellular region is appropriate for this soluble processed-product observation. It does not establish that every intact splice product circulates or that these fragments are confined to matrix.

Summary: The human aortic proteomics source supports extracellular-matrix localization more specifically.

Reason: The original PMID:20551380 Tables I–III identify COIA1_HUMAN in NaCl, guanidine and SDS fractions, with 11, 45 and 6 unique peptides respectively. The actual target rows and tissue-fractionation Methods were inspected. Together with the established basement-membrane role, this supports refining the location to extracellular matrix; the peptide group does not resolve all precursor isoforms or demonstrate isolated tensile mechanics.

Proposed replacements: extracellular matrix

Summary: Extracellular localization agrees with the human venous-ECM proteomics annotation.

Reason: PMID:27068509 studies enriched extracellular fractions of human normal and varicose veins. The exact COL18A1 supplementary target row was not recovered: the authentic supplementary archive contained figure images, not the required workbook. Retain the curator-supported broad location because independent human extracellular and matrix observations establish it; do not claim a personally verified target measurement or disease-dependent change from this paper.

Summary: The combined secreted-protein inference is biologically supported.

Reason: The literal rat donor A0A8I6GLR1 is Col18a1 and carries extracellular-region IDA from PMID:11353854; the combined source also includes the Ensembl product and UniProt secreted-location vocabulary. Human circulating fragments and matrix proteomics independently support the target location. This retains the combined provenance rather than treating it as an isoform-resolved secretion experiment.

Summary: Ortholog-based extracellular localization is supported.

Reason: The authenticated P39061 donor is mouse Col18a1, with experimental extracellular records including PMID:10626789. Its positive source is consistent with human extracellular fragments and matrix detection. The transfer is retained at gene-product scope without equating all mouse constructs and human splice forms.

Summary: The Reactome event supports extracellular localization at its modeled product scope.

Reason: The event represents export of collagen-containing cargo from the ER to the extracellular region; it does not make COL18A1 the secretion machinery. Exact event and nested P39060 entity paths are recorded in the source index.

Summary: The Reactome event supports extracellular localization at its modeled product scope.

Reason: The event locates collagen XVIII and released endostatin extracellularly. Proteases perform release; COL18A1 is not assigned their catalytic activity. Exact event and nested P39060 entity paths are recorded in the source index.

Summary: The Reactome event supports extracellular localization at its modeled product scope.

Reason: The event models extracellular cleavage of collagens XV/XVIII and release of related peptides. The live model contains an implausible endCoordinate 11754 for one P39060 fragment; this coordinate is recorded as a source-model limitation, not copied into product mapping. Exact event and nested P39060 entity paths are recorded in the source index.

Summary: The Reactome event supports extracellular localization at its modeled product scope.

Reason: The event models extracellular endostatin degradation by cathepsins. The fragment is the substrate, not the enzyme executing degradation. Exact event and nested P39060 entity paths are recorded in the source index.

Summary: The Reactome event supports extracellular localization at its modeled product scope.

Reason: The event models extracellular association of an endostatin dimer with laminin-111. It supports a processed-product location and does not imply every precursor is a dimer. Exact event and nested P39060 entity paths are recorded in the source index.

Summary: The Reactome event supports extracellular localization at its modeled product scope.

Reason: The event models extracellular endostatin association with integrin complexes. Its location is retained without creating a new receptor-binding mechanism from event membership alone. Exact event and nested P39060 entity paths are recorded in the source index.

Summary: Refine collagen-trimer membership to the collagen-XVIII homotrimer.

Reason: The original Knobloch-syndrome reference is preserved, but its available abstract is not a direct trimer-assembly assay. Independent PMID:19631658 demonstrates trimerization of an isolated human collagen-XVIII carboxy-terminal domain, with murine full-NC1 oligomerization assessed separately; the collagen-family synthesis in PMID:21421911 supports the type-specific complex. Refine to collagen type XVIII trimer without attributing the later experiments to the original clinical paper.

Proposed replacements: collagen type XVIII trimer

Summary: Basement-membrane localization is a defining collagen-XVIII property.

Reason: The subcellular-location mapping is consistent with human tissue localization and mouse basement-membrane immunostaining/knockout controls. Mouse loss alters basement-membrane structure and vitreoretinal anchoring. Retain the location without claiming that every endostatin fragment remains attached to basement membrane. Independent PMID:9738008 provides chick basal-lamina and heparan-sulfate evidence; it is supporting ortholog context, not an additional human experiment.

Summary: ER-lumen residence reflects precursor biosynthesis.

Reason: The current ARBA rule has a literal target match through IPR008160, IPR013320 and the primate condition. Signal-directed secretion and procollagen assembly make ER-lumen residence reasonable during biosynthesis, but the established extracellular matrix function is the core location. This is not an additional ER enzymatic activity. Literal rule and target-feature evidence: file:human/COL18A1/COL18A1-source-evidence.json, arba_current_target_conditions.

Summary: The modeled ER location is a biosynthetic stage.

Reason: Triple-helix formation occurs in the ER; the collagen chain is the chain being assembled, while the named processing factors perform their own catalytic steps. Retain the compartment at that stage, not as the mature protein's principal site of matrix function.

Summary: The modeled ER location is a biosynthetic stage.

Reason: The secretion reaction contains ER collagen cargo before extracellular export. Retain the compartment at that stage, not as the mature protein's principal site of matrix function.

Summary: The modeled ER location is a biosynthetic stage.

Reason: Association of procollagen type XVIII explicitly models the precursor homotrimer in the ER. Retain the compartment at that stage, not as the mature protein's principal site of matrix function.

Summary: Vision depends on collagen-XVIII-supported ocular architecture.

Reason: The original human PMID:10942434 abstract reports a segregating splice-site defect in Knobloch syndrome and expression of the affected short form in retina. Mouse PMID:11927538 independently connects loss to vitreoretinal and vascular abnormalities. These preserve a contextual visual process without implying that collagen XVIII executes phototransduction or that the historical transcript designation identifies every current product.

Summary: Some endostatin preparations inhibit endothelial proliferation.

Reason: The original PMID:9008168 explicitly reports inhibitory endostatin, including an E. coli-derived preparation administered as a nonrefolded suspension. In contrast, PMID:9459295 reports no antiproliferative activity for purified native circulating human endostatin on the tested endothelial cells, and other NC1 oligomers have distinct effects. Retain the positive original result at preparation-specific scope rather than reject it or universalize it to all intact precursors and fragments.

Summary: The xenobiotic-response transfer has a positive rat endostatin context.

Reason: The authenticated rat Col18a1 record carries IEP from PMID:11353854. Ticlopidine treatment changes serum/platelet-associated endostatin, and anti-endostatin neutralization restores HUVEC proliferation and reduces the serum-associated apoptotic effect in the tested assay. This supports a contextual response to the administered drug, not drug metabolism or an autonomous xenobiotic sensor, and not a direct human treatment experiment.

Summary: Broad anatomical morphogenesis is consistent with the protein’s structural and processed-product effects.

Reason: The current ARBA00026583 conditions match the target through three exact FunFam entries. Clinical retinal/neural-closure observations and the NC1 morphogenesis experiments support the broad biology, but it adds a contextual outcome rather than a second molecular core. No historical rule error is inferred. Literal rule and target-feature evidence: file:human/COL18A1/COL18A1-source-evidence.json, arba_current_target_conditions.

Summary: The original clinical source supports organ morphogenesis in a disease context.

Reason: PMID:10942434 links COL18A1 disruption with severe ocular abnormalities and variable occipital defects. Its available abstract discusses primary or secondary effects on proliferation/migration without resolving a separate developmental mechanism. Retain the broad organ-level process as non-core; do not infer that the collagen chain executes an enzymatic developmental pathway.

Summary: Retain the inherited collagen structural activity conferring tensile strength.

Reason: The official term explicitly includes extracellular-matrix collagens. Collagen XVIII forms homotrimers and contributes to basement-membrane architecture; a nonfibrillar designation does not refute tensile structural participation. Same-role collagen XV carries this term, and network-forming collagen IV has direct experimental support. Historical PTN placement was not reconstructed, but no target-specific loss was established; donor number and absence of an isolated force assay are not reasons to reject the inference.

Summary: Retain the collagen structural classification, distinguished from the proteomics measurement.

Reason: The exact human aortic target is present in Tables I–III of the original source. The RCA row is a reviewed computational/class-level structural assignment, not a direct tensile-force assay. The official collagen convention, target trimer assembly and basement-membrane role support retention. Source-specific peptide/protein detection and biomechanical interpretation are kept separate; no full-product mechanical equivalence is asserted.

Summary: Retain the collagen structural classification, distinguished from the proteomics measurement.

Reason: The genuine full variant describes human endothelial proteome and deposited-matrix profiling; its exact COL18A1 supplementary entry was not independently recovered. The RCA row is a reviewed computational/class-level structural assignment, not a direct tensile-force assay. The official collagen convention, target trimer assembly and basement-membrane role support retention. Source-specific peptide/protein detection and biomechanical interpretation are kept separate; no full-product mechanical equivalence is asserted.

Summary: Retain the collagen structural classification, distinguished from the proteomics measurement.

Reason: The authentic supplement identifies the COL18A1 group in complete and matrisome sheets, including P39060 and legacy/isoform accessions. The RCA row is a reviewed computational/class-level structural assignment, not a direct tensile-force assay. The official collagen convention, target trimer assembly and basement-membrane role support retention. Source-specific peptide/protein detection and biomechanical interpretation are kept separate; no full-product mechanical equivalence is asserted.

Summary: Retain the collagen structural classification, distinguished from the proteomics measurement.

Reason: The genuine full variant describes human atrial ECM/glycoproteomics; the exact COL18A1 target supplement remains uninspected. The RCA row is a reviewed computational/class-level structural assignment, not a direct tensile-force assay. The official collagen convention, target trimer assembly and basement-membrane role support retention. Source-specific peptide/protein detection and biomechanical interpretation are kept separate; no full-product mechanical equivalence is asserted.

Summary: Retain the collagen structural classification, distinguished from the proteomics measurement.

Reason: The genuine full variant describes the ECM proteomics workflow; the linked target workbook was blocked by an HTML challenge and was not a recovered spreadsheet. The RCA row is a reviewed computational/class-level structural assignment, not a direct tensile-force assay. The official collagen convention, target trimer assembly and basement-membrane role support retention. Source-specific peptide/protein detection and biomechanical interpretation are kept separate; no full-product mechanical equivalence is asserted.

Summary: Collagen XVIII contributes structure to extracellular-matrix organization.

Reason: The chain is part of the assembled matrix rather than merely an input consumed during its processing. Direct isolated-domain trimerization and mouse basement-membrane/vitreoretinal abnormalities support structural participation. The broad IBA is retained without claiming that all extant donor evidence reconstructs the historical PTN002770636 placement or that proteases acting on collagen confer protease activity on it.

Summary: Collagen-XVIII trimer membership is supported independently of current rule availability.

Reason: ARBA00098128 could not be retrieved as a current rule; that is not proof that the historical electronic assignment was wrong. The specific complex is supported by the collagen-family reference and the human trimerization-domain experiment, with murine full-NC1 oligomerization distinguished. Literal rule and target-feature evidence: file:human/COL18A1/COL18A1-source-evidence.json, arba_current_target_conditions.

Summary: The family synthesis identifies the collagen-XVIII homotrimer.

Reason: PMID:21421911 is a review, not a newly performed structural experiment. Its class assignment agrees with direct human isolated-domain trimerization in PMID:19631658. Retain the specific complex while distinguishing complete precursor assembly from the behavior of cleaved endostatin monomers.

Summary: Human aortic extracellular fractions directly contain collagen XVIII.

Reason: The exact COIA1_HUMAN rows of PMID:20551380 Tables I–III were inspected with the extraction Methods and table footnotes. The peptide/protein-level localization supports extracellular matrix; the original source qualifier is preserved and no isoform-resolved or mechanical assay is inferred.

Summary: Human embryonic-stem-cell-derived matrix contains the COL18A1 group.

Reason: Original PMID:23658023 Table 1 lists COL18A1 at normalized spectral count 5.83 in HUES1 ECM, with ND in six feeder columns. Methods describe feeder-free HUES1 matrix preparation and separate species databases. This is positive human matrix evidence, not a mouse-feeder transfer; ND entries are not generalized absence claims.

Summary: Retain the established extracellular-matrix location with source-specific access limits.

Reason: PMID:23979707 provides human endothelial deposited-matrix profiling. The exact COL18A1 supplemental target row remains uninspected despite genuine main-text access/recovery attempts. Independent human target tables and the known basement-membrane protein establish the location, so the curated HDA assertion is retained without claiming a new target-specific check or a function inferred from a title.

Summary: Human colon/liver ECM proteomics detects the COL18A1 protein group.

Reason: PMID:25037231 provides human colon/liver tissue ECM enrichment. The exact COL18A1 protein group was verified in both supplement sheets at row 110; these are duplicate presentations of one group, not independent experiments. Its P39060 and legacy/isoform accession group supports extracellular-matrix localization, without resolving each precursor isoform or attributing condition-specific changes or tensile-force measurements to the group.

Summary: Retain the established extracellular-matrix location with source-specific access limits.

Reason: PMID:27559042 provides human atrial ECM/glycoproteomic profiling. The exact COL18A1 supplemental target row remains uninspected despite genuine main-text access/recovery attempts. Independent human target tables and the known basement-membrane protein establish the location, so the curated HDA assertion is retained without claiming a new target-specific check or a function inferred from a title.

Summary: Retain the established extracellular-matrix location with source-specific access limits.

Reason: PMID:28675934 provides an ECM isolation and mass-spectrometry workflow. The exact COL18A1 supplemental target row remains uninspected despite genuine main-text access/recovery attempts. Independent human target tables and the known basement-membrane protein establish the location, so the curated HDA assertion is retained without claiming a new target-specific check or a function inferred from a title.

Summary: The inherited extracellular-matrix location is supported for the structural protein.

Reason: The positive target matrix observations and collagen-XVIII basement-membrane biology support the inherited location. All literal source entities and the existing qualifier are retained. The source node is a phylogenetic assertion, not a count of independent binding or localization experiments; its historical placement remains separately unresolved.

Summary: Extracellular-matrix association is retained without overstating the original species/target verification.

Reason: The PMID:22261194 abstract describes a porcine ischemia/reperfusion model and explicitly states that selected findings were subsequently confirmed in human left ventricle. It does not identify COL18A1 as one of those validated human targets, and the target-level full evidence was not recovered. Independent human ECM measurements make the location sound; retain the TAS assertion and original qualifier while leaving the specific citation-target link unverified.

Summary: Cell-population proliferation regulation is preparation and context dependent.

Reason: The current ARBA00027546 target match comprises the three exact FunFam conditions and the non-fungal condition. Positive endostatin and NC1 experiments support regulatory effects, but native human circulating material can lack the tested antiproliferative effect, and embryonic contexts can show positive proliferation. Retain the broad regulation term without imposing one sign or one activity on every precursor. Literal rule and target-feature evidence: file:human/COL18A1/COL18A1-source-evidence.json, arba_current_target_conditions.

Summary: Hydrostatic-pressure response is supported in rat tendon-cell context.

Reason: The literal rat source carries IEP from PMID:12798059. Its abstract reports increased endostatin under intermittent hydrostatic pressure and inhibition of HUVEC proliferation by the resulting conditioned medium. The abstract does not describe source-specific neutralization, so the medium effect is not uniquely assigned to endostatin. Retain the response annotation as contextual ortholog evidence, without an autonomous pressure-sensor or direct human pressure-response claim.

Summary: The specific exosomal COL18A1 assignment remains unverified.

Reason: The original source studies urinary exosomes, and genuine main-text material was read. However, the COL18A1 target row/peptide evidence in the relevant supplement was not recovered. Extracellular secretion or matrix localization alone does not settle vesicle association. Preserve the positive HDA source assertion as undecided; there is no evidence here for contamination, misattribution, or absence from exosomes.

Summary: The specific exosomal COL18A1 assignment remains unverified.

Reason: The original source studies human parotid-gland exosomes, and genuine main-text material was read. However, the COL18A1 target row/peptide evidence in the relevant supplement was not recovered. Extracellular secretion or matrix localization alone does not settle vesicle association. Preserve the positive HDA source assertion as undecided; there is no evidence here for contamination, misattribution, or absence from exosomes.

Summary: The specific exosomal COL18A1 assignment remains unverified.

Reason: The original source studies expressed-prostatic-secretion exosome preparations, and genuine main-text material was read. However, the COL18A1 target row/peptide evidence in the relevant supplement was not recovered. Extracellular secretion or matrix localization alone does not settle vesicle association. Preserve the positive HDA source assertion as undecided; there is no evidence here for contamination, misattribution, or absence from exosomes.

Summary: The elastic-fibre association is supported at processed endostatin scope.

Reason: The current ARBA00084594 conditions do not match the inspected target features, but that cannot reconstruct a historical rule failure. The authentic publisher PDF of PMID:23962539 Table 1 lists endostatin, explicitly the C-terminal collagen-XVIII product, with vascular elastic fibres. Retain this contextual association without extending it to every intact precursor or treating it as elastic-fibre assembly activity. Literal rule and target-feature evidence: file:human/COL18A1/COL18A1-source-evidence.json, arba_current_target_conditions.

Summary: The cited review places endostatin alongside vascular elastic fibres.

Reason: Original PMID:23962539 Table 1 was inspected in the genuine publisher PDF: its entity is endostatin, the C-terminus of collagen XVIII, and its location is vascular elastic fibres. This positive review-level statement supports the TAS location at fragment scope, not whole-precursor/all-isoform localization or an independently performed assembly process. The protected normal cache remains abstract-only.

Summary: Broad animal development is consistent with structural tissue consequences.

Reason: The current ARBA00098139 rule could not be retrieved; no historical mapping failure is inferred. Human ocular/occipital clinical findings and mouse retinal/basement-membrane development support a broad contextual developmental role, while the molecular core remains matrix structure. Preserve the original electronic source and distinguish it from a directly measured universal developmental activity. Literal rule and target-feature evidence: file:human/COL18A1/COL18A1-source-evidence.json, arba_current_target_conditions.