COL3A1

Summary: Retain the inherited embryonic-development context.

Reason: The exact mouse donor carries IMP evidence from Col3a1 genetic complementation. The original Tsk2 study distinguishes a gain-of-function C33S allele from null alleles and reports failure of compound mutant/null animals to survive. Collagen supplies a structural matrix component during development; this supports contextual involvement without assigning an embryo-patterning signal or treating every mutant phenotype as a normal collagen activity.

Summary: Retain neuron migration as a ligand-and-matrix context.

Reason: Current mouse donor evidence and the original cortical studies show abnormal neuronal positioning after Col3a1 loss. Collagen III also performs receptor-ligand work that limits migration through GPR56, rather than merely being expressed near migrating neurons. Preserve the broad inherited process and its species boundary.

Summary: Retain tissue homeostasis through structural matrix maintenance.

Reason: The current mouse donor has experimental dermal/vascular evidence in addition to an electronic rule. The complete available transgenic-mouse abstract reports disturbed fibrillogenesis and tissue integrity. This is compatible with matrix maintenance, not a separate homeostatic sensor or a claim that all alternative products were tested.

Summary: Retain the positive interaction with MMP12.

Reason: The complete original abstract directly reports recognition and cleavage of human skin type-III collagen by the MMP12 catalytic domain. Protease binding is already an informative partner class. Collagen is the bound and cleaved substrate; neither protease activity nor autonomous proteolysis is assigned to COL3A1. No exact target/source pair was returned by the selected IntAct/MINT exports, which does not negate the primary result.

Summary: The precise chondrocyte-differentiation allocation remains unresolved.

Reason: The current mouse IMP donor points to a study of collagen fibril architecture and biomechanics in postnatal cartilage and meniscus. The actual selected Methods/Results demonstrate structural matrix work and altered mechanical properties, but not acquisition of chondrocyte identity or specialized cell features. Structural participation in differentiation is possible; the measured matrix phenotype alone does not settle this exact process. Preserve the original source rather than treating the positive experiment as disproved.

Summary: Retain integrin binding as an established secondary interaction.

Reason: The complete original abstract reports adhesion to native collagen I/III under static and flow conditions, with alpha2beta1 and von Willebrand factor requirements. It supports the informative integrin-binding assertion. The full reagent/construct Methods were not recovered; no exact current splice-product assignment or autonomous platelet receptor activity is inferred.

Summary: Retain integrin binding in the human matrix-organization context.

Reason: The available original human fibroblast abstract identifies COL3A1-mutant cells, disorganized collagen/fibronectin matrix and changed integrin recruitment. This is concordant with independently demonstrated collagen III-integrin adhesion. Preserve the NAS assertion without describing the mutant-cell study as a purified binding measurement.

Summary: Specify the collagen tensile-strength structural class.

Reason: The InterPro-based collagen assignment is concordant with the independently established fibrillar matrix role. The current GO:0030020 comment explicitly includes extracellular-matrix collagens. This refinement uses that class convention and observed structural matrix contribution; it does not claim an isolated human chain tensile test or activity of every splice product.

Proposed replacements: extracellular matrix structural constituent conferring tensile strength

Summary: Specify the collagen tensile-strength structural class.

Reason: The original human exon-skipping study distinguishes secreted mutant homotrimers that fail matrix incorporation from intracellularly retained mixed trimers. The current GO:0030020 comment explicitly includes extracellular-matrix collagens. This refinement uses that class convention and observed structural matrix contribution; it does not claim an isolated human chain tensile test or activity of every splice product.

Proposed replacements: extracellular matrix structural constituent conferring tensile strength

Summary: Specify the collagen tensile-strength structural class.

Reason: The original human dermal study reports altered collagen fibril size and distribution with COL3A1 variants. The current GO:0030020 comment explicitly includes extracellular-matrix collagens. This refinement uses that class convention and observed structural matrix contribution; it does not claim an isolated human chain tensile test or activity of every splice product.

Proposed replacements: extracellular matrix structural constituent conferring tensile strength

Summary: Represent the experimentally demonstrated LAIR1 ligand activity.

Reason: Actual original Methods/Results test purified human collagen III with LAIR1 binding/blocking controls, receptor-dependent reporter activation and ITIM-dependent inhibition. These measurements establish a change in receptor activity, satisfying GO:0048018 beyond generic physical association. The physiological inhibitory context is preserved separately; receptor clustering is not evidence of covalent peptide cross-linking.

Proposed replacements: receptor ligand activity

Summary: Specify binding to the platelet-adhesion molecule von Willebrand factor.

Reason: The complete original abstract identifies a high-affinity site using synthetic triple-helical collagen-III peptides and tests platelet adhesion. The official VWF function is to bridge subendothelial collagen and the platelet-surface receptor complex, supporting the cell-adhesion-molecule class of GO:0050839. Preserve the positive pair; do not infer that collagen itself is a platelet receptor, that this is the only VWF adhesion mechanism, or that a current isoform was sequence-authenticated. Full original experimental text was not recovered. The freshly checked parent definition requires binding to a cell adhesion molecule, with no cell-surface restriction. Its child GO:0098631 explicitly requires a cell-surface mediator and does not fit soluble VWF; that child restriction is not imposed on GO:0050839. Current official P04275 FUNCTION and experimental cell-adhesion annotations support the adhesive partner class. This is a positive class refinement, not an inference that VWF is a membrane receptor or that an existing GOA precedent is required.

Proposed replacements: cell adhesion molecule binding

Summary: Retain the positive interaction with mosquito aegyptin.

Reason: Two literal MINT ELISA records map human COL3A1 to the Aedes aegypti product O01949; one records a collagen binding-associated segment. Actual primary text supports aegyptin binding collagen III and the VWF-binding peptide. The soluble collagen reagent species is not specified in the inspected Methods, and mixed Horm collagen used elsewhere is horse tendon. Preserve the curated human mapping with that limit, without asserting a COL3A1 enzyme-inhibitor activity or transferring aegyptin pharmacology to collagen.

Summary: Specify the collagen–ADGRG1 interaction as receptor-ligand activity.

Reason: The preserved human IPI source names ADGRG1/GPR56; its available clinical abstract does not expose the exact human binding assay. Independent original PMID:21768377 main Results and Methods show collagen-III-triggered RhoA activation in mouse systems, loss of that response after GPR56 knockdown or deletion, and inhibition by dominant-negative G-alpha13. This supports the functional GO:0048018 receptor-ligand activity refinement rather than only the GPCR partner class. Retain the original human source and its assay-access limit; the commercial collagen reagent and human Fc tag are not treated as sequence-authenticated direct human affinity evidence. No new annotation or all-isoform claim is added.

Proposed replacements: receptor ligand activity

Summary: Specify binding to the platelet-adhesion molecule von Willebrand factor.

Reason: The complete original abstract reports direct VWF binding to collagen I/III despite a distinct VWF contribution to adhesion on subendothelium. The official VWF function is to bridge subendothelial collagen and the platelet-surface receptor complex, supporting the cell-adhesion-molecule class of GO:0050839. Preserve the positive pair; do not infer that collagen itself is a platelet receptor, that this is the only VWF adhesion mechanism, or that a current isoform was sequence-authenticated. Full original experimental text was not recovered. The freshly checked parent definition requires binding to a cell adhesion molecule, with no cell-surface restriction. Its child GO:0098631 explicitly requires a cell-surface mediator and does not fit soluble VWF; that child restriction is not imposed on GO:0050839. Current official P04275 FUNCTION and experimental cell-adhesion annotations support the adhesive partner class. This is a positive class refinement, not an inference that VWF is a membrane receptor or that an existing GOA precedent is required.

Proposed replacements: cell adhesion molecule binding

Summary: Retain extracellular detection in human vein proteomics.

Reason: The original human vein Methods and Table 1 identify collagen alpha-1(III) in the NaCl-extracted extracellular fraction. The exact target row reports 13.7 versus 3.5 mean total-ion-current units in normal versus varicose veins, without isoform resolution. This supports the source location; it is not a target-specific force assay or evidence that the protein catalyzes remodeling.

Summary: Retain secretion into the extracellular region.

Reason: The complete original human variant abstract explicitly separates secreted mutant homotrimers from retained mixed trimers and matrix incorporation. Extracellular secretion remains valid even when a mutant product fails stable matrix incorporation.

Summary: Retain the extracellular location of structural collagen III.

Reason: The original human dermal variant abstract describes collagen fibril defects in connective tissue. The preserved NAS location is independently corroborated by direct human matrix proteomics and is not restricted to one mature product or disease state.

Summary: Retain the modeled extracellular region location.

Reason: The exact Reactome event R-HSA-1454757 (Gelatin degradation by MMP1, 2, 3, 7, 8, 9, 12, 13) has 72 enumerated membership/component routes to P02461. A representative route resolves 3x4Hyp-3Hyp-5Hyl-COL3A1(154-948) [extracellular region]. The curated event and compartment support this location. Collagen can be a structural participant, ligand or processing substrate in different events; this CC assertion does not assign the enzyme, trafficking machinery or receptor activity to the chain. Set membership is not a new target-specific assay.

Summary: Retain the modeled extracellular region location.

Reason: The exact Reactome event R-HSA-1474210 (Collagen type III degradation by MMP14) has 108 enumerated membership/component routes to P02461. A representative route resolves 3x4Hyp-3Hyp-GalHyl-COL3A1 [extracellular region]. The curated event and compartment support this location. Collagen can be a structural participant, ligand or processing substrate in different events; this CC assertion does not assign the enzyme, trafficking machinery or receptor activity to the chain. Set membership is not a new target-specific assay.

Summary: Retain the modeled extracellular region location.

Reason: The exact Reactome event R-HSA-1474213 (Collagen type III degradation by MMP1,8,9,13) has 108 enumerated membership/component routes to P02461. A representative route resolves 3x4Hyp-3Hyp-GalHyl-COL3A1 [extracellular region]. The curated event and compartment support this location. Collagen can be a structural participant, ligand or processing substrate in different events; this CC assertion does not assign the enzyme, trafficking machinery or receptor activity to the chain. Set membership is not a new target-specific assay.

Summary: Retain the modeled extracellular region location.

Reason: The exact Reactome event R-HSA-1474266 (Formation of collagen fibrils) has 72 enumerated membership/component routes to P02461. A representative route resolves 3x4Hyp-3Hyp-GalHyl-COL3A1 [extracellular region]. The curated event and compartment support this location. Collagen can be a structural participant, ligand or processing substrate in different events; this CC assertion does not assign the enzyme, trafficking machinery or receptor activity to the chain. Set membership is not a new target-specific assay.

Summary: Retain the modeled extracellular region location.

Reason: The exact Reactome event R-HSA-2002428 (Removal of fibrillar collagen N-propeptides) has 72 enumerated membership/component routes to P02461. A representative route resolves 3x4Hyp-3Hyp-GalHyl-COL3A1(24-1466) [extracellular region]. The curated event and compartment support this location. Collagen can be a structural participant, ligand or processing substrate in different events; this CC assertion does not assign the enzyme, trafficking machinery or receptor activity to the chain. Set membership is not a new target-specific assay.

Summary: Retain the modeled extracellular region location.

Reason: The exact Reactome event R-HSA-2002440 (Removal of fibrillar collagen C-propeptides) has 72 enumerated membership/component routes to P02461. A representative route resolves 3x4Hyp-3Hyp-GalHyl-COL3A1(154-1466) [extracellular region]. The curated event and compartment support this location. Collagen can be a structural participant, ligand or processing substrate in different events; this CC assertion does not assign the enzyme, trafficking machinery or receptor activity to the chain. Set membership is not a new target-specific assay.

Summary: Retain the modeled extracellular region location.

Reason: The exact Reactome event R-HSA-2089971 (Secretion of collagens) has 72 enumerated membership/component routes to P02461. A representative route resolves 3x4Hyp-3Hyp-GalHyl-COL3A1(24-1466) [extracellular region]. In this secretion event the ER-lumen input is distinct from the extracellular output selected here; the output route supports this location. Collagen can be a structural participant, ligand or processing substrate in different events; this CC assertion does not assign the enzyme, trafficking machinery or receptor activity to the chain. Set membership is not a new target-specific assay.

Summary: Retain the modeled extracellular region location.

Reason: The exact Reactome event R-HSA-2213201 (Formation of collagen fibres) has 72 enumerated membership/component routes to P02461. A representative route resolves 3x4Hyp-3Hyp-GalHyl-COL3A1 [extracellular region]. The curated event and compartment support this location. Collagen can be a structural participant, ligand or processing substrate in different events; this CC assertion does not assign the enzyme, trafficking machinery or receptor activity to the chain. Set membership is not a new target-specific assay.

Summary: Retain the modeled extracellular region location.

Reason: The exact Reactome event R-HSA-2327695 (Collagen types III, IV, V, VI, VIII, IX, XVI bind integrins alpha1beta1 and alpha2beta1) has 72 enumerated membership/component routes to P02461. A representative route resolves 3x4Hyp-3Hyp-GalHyl-COL3A1 [extracellular region]. The curated event and compartment support this location. Collagen can be a structural participant, ligand or processing substrate in different events; this CC assertion does not assign the enzyme, trafficking machinery or receptor activity to the chain. Set membership is not a new target-specific assay.

Summary: Retain the modeled extracellular region location.

Reason: The exact Reactome event R-HSA-2327733 (FN1 binds Collagen types I-V, VII) has 72 enumerated membership/component routes to P02461. A representative route resolves 3x4Hyp-3Hyp-GalHyl-COL3A1 [extracellular region]. The curated event and compartment support this location. Collagen can be a structural participant, ligand or processing substrate in different events; this CC assertion does not assign the enzyme, trafficking machinery or receptor activity to the chain. Set membership is not a new target-specific assay.

Summary: Retain the modeled extracellular region location.

Reason: The exact Reactome event R-HSA-2327738 (DDR1 binds collagen type I, II, III, IV, V, XI fibrils) has 72 enumerated membership/component routes to P02461. A representative route resolves 3x4Hyp-3Hyp-GalHyl-COL3A1 [extracellular region]. The curated event and compartment support this location. Collagen can be a structural participant, ligand or processing substrate in different events; this CC assertion does not assign the enzyme, trafficking machinery or receptor activity to the chain. Set membership is not a new target-specific assay.

Summary: Retain the modeled extracellular region location.

Reason: The exact Reactome event R-HSA-2327909 (DCN binds collagen I, II, III, VI fibrils) has 72 enumerated membership/component routes to P02461. A representative route resolves 3x4Hyp-3Hyp-GalHyl-COL3A1 [extracellular region]. The curated event and compartment support this location. Collagen can be a structural participant, ligand or processing substrate in different events; this CC assertion does not assign the enzyme, trafficking machinery or receptor activity to the chain. Set membership is not a new target-specific assay.

Summary: Retain the modeled extracellular region location.

Reason: The exact Reactome event R-HSA-2396370 (VTN binds collagens II, III and V) has 72 enumerated membership/component routes to P02461. A representative route resolves 3x4Hyp-3Hyp-GalHyl-COL3A1 [extracellular region]. The curated event and compartment support this location. Collagen can be a structural participant, ligand or processing substrate in different events; this CC assertion does not assign the enzyme, trafficking machinery or receptor activity to the chain. Set membership is not a new target-specific assay.

Summary: Retain the modeled extracellular region location.

Reason: The exact Reactome event R-HSA-2465890 (DDR2 binds collagen type I, II, III, V, X fibrils) has 72 enumerated membership/component routes to P02461. A representative route resolves 3x4Hyp-3Hyp-GalHyl-COL3A1 [extracellular region]. The curated event and compartment support this location. Collagen can be a structural participant, ligand or processing substrate in different events; this CC assertion does not assign the enzyme, trafficking machinery or receptor activity to the chain. Set membership is not a new target-specific assay.

Summary: Retain the modeled extracellular region location.

Reason: The exact Reactome event R-HSA-2473584 (Collagen type III degradation by MMP15) has 108 enumerated membership/component routes to P02461. A representative route resolves 3x4Hyp-3Hyp-GalHyl-COL3A1 [extracellular region]. The curated event and compartment support this location. Collagen can be a structural participant, ligand or processing substrate in different events; this CC assertion does not assign the enzyme, trafficking machinery or receptor activity to the chain. Set membership is not a new target-specific assay.

Summary: Retain the modeled extracellular region location.

Reason: The exact Reactome event R-HSA-375151 (Interaction of NCAM1 with collagens) has 72 enumerated membership/component routes to P02461. A representative route resolves 3x4Hyp-3Hyp-GalHyl-COL3A1 [extracellular region]. The curated event and compartment support this location. Collagen can be a structural participant, ligand or processing substrate in different events; this CC assertion does not assign the enzyme, trafficking machinery or receptor activity to the chain. Set membership is not a new target-specific assay.

Summary: Retain the modeled extracellular region location.

Reason: The exact Reactome event R-HSA-382054 (PDGF binds to extracellular matrix proteins) has 72 enumerated membership/component routes to P02461. A representative route resolves 3x4Hyp-3Hyp-GalHyl-COL3A1 [extracellular region]. The curated event and compartment support this location. Collagen can be a structural participant, ligand or processing substrate in different events; this CC assertion does not assign the enzyme, trafficking machinery or receptor activity to the chain. Set membership is not a new target-specific assay.

Summary: Retain the modeled extracellular region location.

Reason: The exact Reactome event R-HSA-5696356 (OSCAR binds collagen and SP-D) has 72 enumerated membership/component routes to P02461. A representative route resolves 3x4Hyp-3Hyp-GalHyl-COL3A1 [extracellular region]. The curated event and compartment support this location. Collagen can be a structural participant, ligand or processing substrate in different events; this CC assertion does not assign the enzyme, trafficking machinery or receptor activity to the chain. Set membership is not a new target-specific assay.

Summary: Retain the modeled extracellular region location.

Reason: The exact Reactome event R-HSA-5696357 (LAIR2 binds collagen) has 72 enumerated membership/component routes to P02461. A representative route resolves 3x4Hyp-3Hyp-GalHyl-COL3A1 [extracellular region]. The curated event and compartment support this location. Collagen can be a structural participant, ligand or processing substrate in different events; this CC assertion does not assign the enzyme, trafficking machinery or receptor activity to the chain. Set membership is not a new target-specific assay.

Summary: Retain the collagen-trimer classification.

Reason: The current mouse donor contains direct collagen-trimer annotations, and independent human type-III chain association establishes the same complex class. Preserve the broader electronic term without manufacturing a separate core or transferring a donor assay to every human alternative product.

Summary: Retain the fibrillar-collagen trimer rule.

Reason: The retrieved ARBA rule predicts this class from a primate FunFam condition. Its full target-specific FunFam activation path was not independently reconstructed. Human type-III collagen nevertheless forms a fibrillar homotrimer, independently supported by structural and dermal evidence; the rule is not the sole basis for acceptance.

Summary: Retain the PAINT type-III homotrimer assignment.

Reason: The target is experimentally grounded for this complex and therefore legitimately appears among the PAINT supporting entities. Current exact-term human and mouse records corroborate the complex class. The historical IBD node and alignment were not reconstructed, so no claim about their placement is added.

Summary: Retain type-III trimer as the assayed collagen class.

Reason: The original MMP12 abstract explicitly assays human skin collagen III and its triple-helical structure. Preserve this complex classification without interpreting protease binding or cleavage as COL3A1 catalytic activity.

Summary: Retain the conserved type-III homotrimer complex.

Reason: The current mouse exact-term donor includes experimental evidence as well as inherited/transfer records. Independent human assembly and secretion evidence supports the target complex; this is not based on counting donors or a self-propagation loop.

Summary: Retain the human type-III collagen complex context.

Reason: The available original abstract reports COL3A1 RNA-interference experiments in human skin fibroblasts. It does not independently determine trimer stoichiometry, but that complex identity is directly established by structural and secretion evidence. Retain the source assertion with the distinction between perturbation and assembly measurement.

Summary: Retain human collagen-III homotrimer assembly.

Reason: The complete available original abstract distinguishes homotrimers assembled from mutant chains from mixed trimers and reports their different secretion and incorporation. This is direct assembly context, without an all-isoform generalization.

Summary: Retain the type-III collagen triple-helical complex.

Reason: The original structural abstract examines a synthetic native-sequence human C-terminal collagen-III peptide containing the cystine knot. It supports triple-helical assembly, while two alternative cystine-knot conformations remain possible. This is not the entire physiological full-length molecule or a current splice-product assay.

Summary: Retain the type-III collagen complex identity.

Reason: The original human splice-site/minigene abstract identifies an exon-skipping COL3A1 product in vascular Ehlers-Danlos syndrome. Its principal experiment is splicing rather than trimer stoichiometry; independent assembly evidence supports the preserved NAS complex assignment.

Summary: Retain the procollagen ER-lumen location.

Reason: The current ARBA rule has a collagen C-terminal domain/Catarrhini condition concordant with the target IPR000885 record. Signal-dependent precursor entry, disulfide-assisted association and the exact Reactome P02461 lumen entities independently support this intracellular biosynthetic location. It is not the location of every mature extracellular chain.

Summary: Retain the modeled endoplasmic reticulum lumen location.

Reason: The exact Reactome event R-HSA-1650808 (Prolyl 4-hydroxylase converts collagen prolines to 4-hydroxyprolines) has 2 enumerated membership/component routes to P02461. A representative route resolves COL3A1(24-1466) [endoplasmic reticulum lumen]. The curated event and compartment support this location. Collagen can be a structural participant, ligand or processing substrate in different events; this CC assertion does not assign the enzyme, trafficking machinery or receptor activity to the chain. Set membership is not a new target-specific assay.

Summary: Retain the modeled endoplasmic reticulum lumen location.

Reason: The exact Reactome event R-HSA-2002460 (P4HB binds Collagen chains) has 2 enumerated membership/component routes to P02461. A representative route resolves COL3A1(24-1466) [endoplasmic reticulum lumen]. The curated event and compartment support this location. Collagen can be a structural participant, ligand or processing substrate in different events; this CC assertion does not assign the enzyme, trafficking machinery or receptor activity to the chain. Set membership is not a new target-specific assay.

Summary: Retain the modeled endoplasmic reticulum lumen location.

Reason: The exact Reactome event R-HSA-2022073 (Procollagen triple helix formation) has 72 enumerated membership/component routes to P02461. A representative route resolves 3x4Hyp-3Hyp-GalHyl-COL3A1(24-1466) [endoplasmic reticulum lumen]. The curated event and compartment support this location. Collagen can be a structural participant, ligand or processing substrate in different events; this CC assertion does not assign the enzyme, trafficking machinery or receptor activity to the chain. Set membership is not a new target-specific assay.

Summary: Retain the modeled endoplasmic reticulum lumen location.

Reason: The exact Reactome event R-HSA-2089971 (Secretion of collagens) has 72 enumerated membership/component routes to P02461. A representative route resolves 3x4Hyp-3Hyp-GalHyl-COL3A1(24-1466) [endoplasmic reticulum lumen]. The curated event and compartment support this location. Collagen can be a structural participant, ligand or processing substrate in different events; this CC assertion does not assign the enzyme, trafficking machinery or receptor activity to the chain. Set membership is not a new target-specific assay.

Summary: Retain the modeled endoplasmic reticulum lumen location.

Reason: The exact Reactome event R-HSA-8944240 (Association of procollagen type III) has 72 enumerated membership/component routes to P02461. A representative route resolves 3x4Hyp-3Hyp-GalHyl-COL3A1(24-1466) [endoplasmic reticulum lumen]. The curated event and compartment support this location. Collagen can be a structural participant, ligand or processing substrate in different events; this CC assertion does not assign the enzyme, trafficking machinery or receptor activity to the chain. Set membership is not a new target-specific assay.

Summary: Retain collagen-supported cell-matrix adhesion.

Reason: The original peptide-mapping abstract links collagen-III VWF binding to platelet adhesion, including flow conditions. The collagen surface performs the adhesive matrix role; this is more than a change in expression. Preserve the source and distinguish the matrix ligand from the platelet receptor and from a quantitative human tissue adhesion assay.

Summary: The exact TGF-beta receptor-pathway participation remains unresolved.

Reason: The complete available source abstract measures increased COL3A1 expression after TGF-beta1 treatment of human bladder smooth-muscle cells. It does not establish which signaling step collagen itself performs. Collagen can influence cells through matrix and ligand functions, but induction alone does not resolve this source-specific pathway assignment. Preserve the positive treatment result; full relevant experimental text was not recovered.

Summary: Retain integrin-mediated signaling in the mutant-matrix context.

Reason: The actual human COL3A1-mutant fibroblast abstract reports altered matrix organization and integrin recruitment, consistent with collagen performing an extracellular ligand/matrix step. The full source was not recovered, so no particular downstream signaling intermediate or purified receptor-activation mechanism is claimed. This is a supported secondary context, not an independent molecular core.

Summary: The precise heart-development assertion remains unresolved.

Reason: The complete original clinical abstract concerns coronary dissection and myocardial infarction in an adolescent with a COL3A1 defect. The collagen structural contribution to vascular integrity is real, but the available source does not establish progression of the heart from formation to maturity. Full experimental text was not recovered; preserve the curator assertion without equating clinical injury with development or denying a possible structural developmental contribution.

Summary: The source establishes radiation-induced collagen-III secretion, not the attributed response function.

Reason: The available original Methods/Results quantify collagen III in conditioned medium from beta-irradiated human Tenon fibroblasts by ELISA, normalized for cell number; increased secretion is genuine. The functional contraction assay instead uses rat-tail collagen I. The Discussion proposes a possible contribution to post-radiation fibrosis, but these experiments do not establish collagen III acting in the radiation response. Given its established matrix role, interpreting this production readout as a direct assay of the attributed response function overstates this source. This is a source-specific evidence-allocation judgment, not a claim that expression-based response annotations are universally invalid or that structural proteins cannot participate. The contextual wound-strength role remains separately supported.

Summary: The reporter evidence leaves the exact response-to-wounding allocation unresolved.

Reason: The current mouse donor is IDA from a Col3a1-promoter EGFP scar reporter. Actual Methods/Results measure transcriptional reporter activity, collagen deposition and corneal scar development; the authors distinguish reporter timing from post-translational protein behavior. These observations do not by themselves determine what work collagen III performs in the asserted response. Preserve the positive signal and the possibility of structural repair participation.

Summary: Mutant-induced fibrotic regulation is over-specific support for a normal gene-expression annotation.

Reason: GO:0010467 describes conversion of a gene sequence into a mature RNA or protein product. The exact mouse IMP records cite Tsk2 promoter-output observations and PMID:25330296, whose original Results compare C33S and wild-type Col3a1 cDNA in Col3a1-null fibroblasts: the mutant increases COL1A1 RNA/protein and a fibrotic expression program. This positive downstream regulatory phenotype is not a demonstrated reference-COL3A1 step in gene-product production. MARK_AS_OVER_ANNOTATED addresses that role and disease-allele extrapolation, not an assertion that extracellular proteins can never regulate gene expression. The original PMID:17960558 full Methods remain unrecovered; its complete abstract and the independent full genetic/complementation study are distinguished. No transcription or gene-expression-regulation NEW is inferred.

Summary: The LAIR1 clustering assay does not support this covalent cross-linking assertion.

Reason: The current GO:0018149 definition requires formation of a covalent link between or within protein chains. Original PMID:16754721 main Results, Figure 5 caption and Methods describe plate-bound collagen III engaging LAIR1, an NFAT reporter and ITIM-dependent inhibition; this is functional receptor clustering, not formation of a covalent protein bond. The biosensor amine/cysteine coupling is experimental immobilization rather than a COL3A1-performed reaction. Remove this source-specific IDA assertion because the assay uses a different meaning of cross-linking. This does not deny collagen III interchain links, its structural contribution to a covalent fibril network, or the separate tissue chemistry in PMID:31655293. No alternative cross-linking process assertion is added.

Summary: Retain cortical layer formation as a contextual ligand/matrix process.

Reason: The current mouse source and original Col3a1-loss study demonstrate cortical disorganization, neuronal overmigration and pial-membrane disruption. Complementary receptor experiments identify a collagen-dependent migration-limiting step. This supports contextual structural/ligand participation in lamination, not an autonomous neural differentiation program or direct human isoform experiment.

Summary: Retain cerebral-cortex development at the orthologous source scope.

Reason: The exact mouse donor and original experiments support a collagen III-GPR56 receptor-ligand axis that limits neuronal migration and preserves cortical organization. Collagen performs ligand work as well as supplying matrix structure. Preserve the original IEA transfer; the commercial collagen reagent in the inspected Methods has no species/catalogue identity specified.

Summary: Retain cerebral-cortex development at the orthologous source scope.

Reason: The exact mouse donor and original experiments support a collagen III-GPR56 receptor-ligand axis that limits neuronal migration and preserves cortical organization. Collagen performs ligand work as well as supplying matrix structure. Preserve the original ISS transfer; the commercial collagen reagent in the inspected Methods has no species/catalogue identity specified.

Summary: Retain the collagen tensile-strength class with exact matrix detection.

Reason: Original human bone-marrow supplementary tables identify COL3A1 in an ECM protein group. The classification follows the official GO collagen convention and independent fibril/biomechanical evidence, not a force measurement in the proteomics experiment. Summary-table MGUS non-detection does not imply universal absence: the individual-sample table has a positive COL3A1 MGUS detection.

Summary: Retain the inherited tensile-strength structural class.

Reason: The PAINT assignment is consistent with the explicit GO collagen convention and target-specific fibril integrity evidence. Supporting collagen-family records span different collagen architectures; their count is not the basis of acceptance. The historical PTN placement was not reconstructed and the target conclusion rests additionally on human dermal and mouse mechanical evidence.

Summary: Retain the collagen tensile-strength classification.

Reason: Actual human colon/liver ECM target rows and extraction Methods are verified; the protein group includes multiple accession/product names. This RCA annotation is a collagen structural classification, not an assertion that the proteomics study measured tensile force. The current GO comment and independent target matrix/biomechanical evidence support the class while preserving the source-access limitation.

Summary: Retain the collagen tensile-strength classification.

Reason: The selected cardiac glycoproteomics cache contains Introduction/Discussion, not complete Methods/Results, and the exact COL3A1 supplement row was not recovered. This RCA annotation is a collagen structural classification, not an assertion that the proteomics study measured tensile force. The current GO comment and independent target matrix/biomechanical evidence support the class while preserving the source-access limitation.

Summary: Retain the collagen tensile-strength classification.

Reason: The selected human cell-derived matrix Introduction/Discussion explicitly identifies collagen III, while the exact supplementary row and full Methods were not independently read. This RCA annotation is a collagen structural classification, not an assertion that the proteomics study measured tensile force. The current GO comment and independent target matrix/biomechanical evidence support the class while preserving the source-access limitation.

Summary: Retain the collagen tensile-strength classification.

Reason: Main protocol-comparison Results/Discussion were read; the exact target supplement could not be authenticated because the saved downloads are HTML challenge responses. This RCA annotation is a collagen structural classification, not an assertion that the proteomics study measured tensile force. The current GO comment and independent target matrix/biomechanical evidence support the class while preserving the source-access limitation.

Summary: Retain the collagen-dependent platelet-activation context.

Reason: The complete original abstract distinguishes platelet adhesion under static and flow conditions and the roles of collagen motifs, integrin and VWF. Collagen supplies the extracellular activation/adhesion surface; it is not itself a platelet signaling enzyme. Preserve the existing secondary process without adding a separate activation core.

Summary: Retain structural participation in matrix organization.

Reason: The current exact mouse evidence includes a gain-of-function fibrotic model and independent Col3a1-reduced cartilage/meniscus experiments. The latter directly measures fibril architecture and matrix mechanics, establishing structural work by collagen beyond necessity alone. Mutant fibrotic signaling is not generalized to every reference product.

Summary: Retain collagen-fibril organization through chain assembly and matrix incorporation.

Reason: Current exact-term collagen-family records include direct target clinical/fibril evidence. Target self-evidence legitimately contributes to PAINT; historical IBD/MSA placement was not reconstructed. Human dermal fibril abnormalities, secretion/incorporation experiments and triple-helical association establish structural participation in organizing fibrils.

Summary: Retain the fibril-organization rule with independent target support.

Reason: The retrieved ARBA rule predicts the process from taxon-conditioned FunFam classes. Its exact target FunFam activation path was not independently reconstructed. Human COL3A1 matrix incorporation and fibril architecture directly corroborate the assigned structural assembly process. No hydroxylase or protease activity is transferred from collagen-processing enzymes.

Summary: Retain collagen fibril organization despite limited original access.

Reason: The available clinical abstract identifies a COL3A1 defect and abnormal collagen behavior in the adolescent patient. The source-specific fibril detail was not fully recovered, but independently established target matrix incorporation and dermal architecture make this function sound. This does not rescue the distinct heart-development term or claim that all clinical findings demonstrate fibril assembly.

Summary: Retain direct structural contribution to dermal fibril organization.

Reason: The original human abstract explicitly reports changes in the size and distribution of dermal collagen fibrils with COL3A1 variants. This is structural work by a collagen chain, not merely expression or consumption by a processing enzyme. Full original Methods were not recovered.

Summary: Retain collagen-fibril organization with the primary-source distinction.

Reason: The original splice-site/minigene study establishes a human COL3A1 variant, while independent human secretion/incorporation and dermal studies establish fibril organization. Preserve the NAS assertion without describing the splice assay as a direct fibril-assembly experiment.

Summary: Retain structural assembly as the basis of fibril organization.

Reason: The original structural abstract studies a synthetic native-sequence human C-terminal collagen-III segment and triple-helical/cystine-knot organization. It provides a molecular assembly premise, complemented by full-tissue fibril evidence. Neither full-length physiological architecture nor a uniquely resolved knot conformation is claimed.

Summary: The lung-development inference remains unresolved at its exact donor source.

Reason: The current mouse donor is a positive IMP annotation from the Tsk2 study. Its complete available abstract explicitly says examined lungs were normal and indistinguishable from controls, while the mutation was then molecularly unidentified; later work identifies a Col3a1 gain-of-function allele. This does not establish that collagen has no lung developmental role, but the original supporting developmental experiment cannot be recovered from the available text. Preserve the positive source as undecided rather than manufacture a NOT assertion.

Summary: Retain extracellular-matrix localization.

Reason: Original human colon/liver ECM fractionation and both exact COL3A1 supplementary target rows are verified. Protein-group accessions do not resolve a particular splice product.

Summary: Retain extracellular-matrix localization.

Reason: The exact cardiac target-table relationship remains unverified: the selected cache exposes Introduction/Discussion, and the original supplementary COL3A1 row was not recovered. Independent direct target ECM evidence supports the location, so the experimental assertion is retained rather than rejected from incomplete access.

Summary: Retain extracellular-matrix localization.

Reason: The actual human cell-derived ECM Discussion identifies collagen III in these matrices. Complete Methods and the exact supplemental row were not independently recovered; this limitation does not negate the positive curator detection or independent matrix localization.

Summary: Retain extracellular-matrix localization.

Reason: Original S2 and five S3 target rows identify COL3A1 in human bone-marrow ECM, with nonuniform detections across samples. The group includes more than one COL3A1 accession/product; no isoform or universal disease-condition claim is made.

Summary: Retain extracellular-matrix localization.

Reason: The available main text describes human breast/omentum ECM proteomics, but the exact COL3A1 table remains uninspected because attempted XLSX responses are HTML. Preserve the target location from the curator and independent direct ECM evidence, with the paper-specific target relationship marked unverified.

Summary: Retain the inherited extracellular-matrix activity/location.

Reason: Current exact-term target and collagen-family records corroborate matrix localization. Target self-evidence is expected experimental grounding, not circularity. The historical PAINT node was not reconstructed; independent human ECM rows and incorporation assays support the target while preserving the original is_active_in qualifier.

Summary: Retain the conserved extracellular-matrix location.

Reason: The mouse donor includes multiple direct location/proteomic records as well as propagated records. Independent human matrix extraction and variant-incorporation evidence supports the target without transferring every donor specimen or asserting a particular alternative product.

Summary: Retain the matrix location established in human fibroblasts.

Reason: The original abstract explicitly describes collagen organization defects in the ECM of COL3A1-mutant human fibroblasts. Matrix incorporation is the relevant observation; changes in integrin recruitment do not erase the structural location. Full original experimental text was not recovered.

Summary: Retain normal collagen-III incorporation into extracellular matrix.

Reason: The original human variant abstract compares matrix incorporation of normal and mutant type-III trimers. Failure of mutant homotrimers to incorporate is evidence about a defective product, not absence of the reference collagen from the matrix.

Summary: Retain the source-scoped extracellular-matrix association.

Reason: The available original abstract explicitly identifies porcine cardiac ECM proteomics. Preserve the original colocalizes_with assertion and the cross-species assay boundary; independent human COL3A1 matrix detections support the location without relabeling the porcine experiment as human.

Summary: The normal-reference TGF-beta1-production contribution remains unresolved.

Reason: The exact mouse IMP donor cites PMID:24820199. Its complete abstract and independently accessed original Figure 5 caption report increased total TGF-beta1 and TGF-beta1-positive cells before Tsk2 fibrosis. Original full Methods were not recovered. Independent full PMID:25330296 establishes the C33S gain-of-function allele and increased Tgfb1-associated fibrotic output, but it does not separate amplified normal COL3A1-dependent production from a mutant-specific pathological pathway. GO:0032905 includes biosynthesis or secretion following a stimulus; a matrix-mediated contribution is biologically possible and is not excluded by extracellular location. Retain UNDECIDED for this unresolved source-to-reference-protein interpretation, not because the later full paper is unavailable or because only an intrinsic synthetic enzyme could participate.

Summary: The exact cytokine-response participation remains unresolved.

Reason: TGF-beta1 increases collagen-III expression in human bladder smooth-muscle cells. These are positive changes in the amount of the gene product. The accessible original abstract does not identify a response step performed by collagen itself, and relevant full text was not recovered. This judgment does not require collagen to be a signaling enzyme or deny possible matrix-mediated response work.

Summary: The exact cytokine-response participation remains unresolved.

Reason: Interferon-gamma changes collagen transcription and RNA stability in human lung fibroblasts. These are positive changes in the amount of the gene product. The accessible original abstract does not identify a response step performed by collagen itself, and relevant full text was not recovered. This judgment does not require collagen to be a signaling enzyme or deny possible matrix-mediated response work.

Summary: Retain the collagen-dependent Rho-signaling context.

Reason: Actual original assays support GPR56- and Galpha12/13-dependent RhoA activation by collagen III, with receptor and competition controls. Collagen performs the ligand step, so this is more than a perturbation phenotype. Preserve the original mouse-to-human IEA scope and the unspecified collagen-reagent species; no collagen GTPase or exchange-factor activity is inferred.

Summary: Retain the collagen-dependent Rho-signaling context.

Reason: Actual original assays support GPR56- and Galpha12/13-dependent RhoA activation by collagen III, with receptor and competition controls. Collagen performs the ligand step, so this is more than a perturbation phenotype. Preserve the original mouse-to-human ISS scope and the unspecified collagen-reagent species; no collagen GTPase or exchange-factor activity is inferred.

Summary: Retain organismal growth as a contextual matrix phenotype.

Reason: The exact mouse source reports reduced body length/weight in Tsk2 animals. Later work identifies the Col3a1 gain-of-function allele. A structural matrix contribution to organism size is plausible and the curator process is retained as contextual; the phenotype is not promoted to a normal growth-factor activity or a separate core. Full original growth experiments were not recovered.

Summary: Retain aortic development in the structural tissue context.

Reason: The exact mouse sources show aortic matrix/ultrastructural defects, altered mechanical integrity and susceptibility to dissection. Actual primary ultrastructure connects collagen deficiency to tissue architecture. Retain this structural developmental context while distinguishing adult integrity/challenge experiments from embryonic patterning and without assigning a vascular signaling mechanism.

Summary: Retain structural participation in postnatal limb-joint organization as non-core.

Reason: This mouse IMP transfer cites PMID:31655293, not the Tsk2 gain-of-function studies. Actual Methods/Results and Discussion examine normal Col3a1 heterozygous reduction during postnatal joint maturation: cartilage and meniscus show altered collagen-fibril architecture and mechanical properties. COL3A1 contributes the matrix structure being organized, rather than being only a passive pathology marker. GO:0036022 covers generation and organization of limb-joint anatomical structures and is not restricted to embryonic patterning or gross joint shape. Keep the contextual process as non-core at the measured postnatal matrix-organization scope. Normal gross anatomy or subchondral parameters in the examined young male mice are not universal negative evidence, and proposed molecular details in the Discussion remain distinguished from direct observations.

Summary: The wound-healing process assignment is unresolved by this marker study.

Reason: The complete available original abstract measures type-I/III procollagen propeptides in human chronic-ulcer fluid, including growth-hormone treatment and correlations with healing. A biomarker increase is genuine but does not itself identify work performed by collagen III in restoring tissue integrity. Its known structural role makes participation plausible; full original evidence for that role in these wounds was not recovered.

Summary: Retain wound healing through the established matrix-strength contribution.

Reason: The actual Tenon-fibroblast paper discusses collagen I/III deposition as a source of wound strength and measures collagen-III secretion after irradiation. Retain the NAS contextual structural role, supported by independent matrix-integrity evidence. The rat-tail collagen-I contraction lattice is not a COL3A1 contraction assay, and the radiation-response assertion is separately marked over-annotated at this source's production-assay scope.

Summary: Retain inherited skin development as a matrix context.

Reason: The PAINT assignment includes current direct target dermal evidence. Collagen III contributes structure to the developing dermis, with target variants altering fibril architecture; the target appearing as its own evidence is not circular. Historical node placement was not reconstructed, and no epidermal patterning activity is asserted.

Summary: Retain the mouse skin-development context with allele limits.

Reason: The current donor records include Tsk2 fibrosis and its later identification as a Col3a1 C33S gain-of-function allele. Collagen III supplies dermal matrix, but the timing and signaling abnormalities of this mutant are not generalized as physiological activity of every human product.

Summary: Retain structural contribution to human skin development.

Reason: The original human COL3A1 variant abstract reports abnormal dermal collagen fibrils. This supports the retained developmental matrix context, not a separate skin-patterning signal. Full original Methods were not recovered.

Summary: Retain the positive SMAD-binding transfer.

Reason: The complete original mouse fetal-orofacial abstract reports yeast two-hybrid interactions and GST-SMAD MH2 pull-downs with collagens I, III and V. This is actual protein interaction evidence, not merely a SMAD-binding DNA promoter sequence. Preserve the current mouse donor and the informative SMAD class; native compartment/processing and the particular human product remain unresolved.

Summary: Tsk2-associated fibroblast expansion does not yet resolve the transferred normal proliferation role.

Reason: The exact mouse IMP donor cites PMID:17960558, whose complete available abstract reports more collagen-expressing dermal fibroblasts per unit area. Its full Methods remain unrecovered. Later full PMID:25330296 assigns Tsk2 to Col3a1 C33S gain-of-function and measures fibrotic RNA/protein output; it does not itself resolve fibroblast multiplication versus altered recruitment, survival or tissue organization in the earlier count. GO:0048144 requires expansion by multiplication or reproduction. Retain UNDECIDED without denying positive expansion, assuming that extracellular matrix cannot affect proliferation, or treating gain-of-function as proof of a wholly novel rather than amplified normal activity.

Summary: The source shows excess elastic fibres, but the normal assembly contribution remains unresolved.

Reason: The exact mouse IMP donor cites PMID:24820199. Its complete abstract preserves the early increase in dermal elastic fibres and the finding that removing skin elastic fibres does not abolish tight skin or fibrosis. Full original Methods remain unrecovered. Independent PMID:25330296 identifies a Col3a1 C33S gain-of-function allele; this does not decide whether the excess fibres reflect amplification of normal matrix assembly or a mutant-specific downstream response. GO:0048251 concerns assembly of recoiling ECM fibres. UNDECIDED preserves that biological uncertainty and the positive phenotype; lack of dependence of tight skin on elastic fibres is not evidence against fibre assembly itself.

Summary: Retain direct platelet-derived growth factor binding.

Reason: The complete original abstract reports PDGF isoforms binding collagen III among other collagen types, including radioligand binding and bioavailability effects. This supports the informative binding class. It does not make collagen a PDGF receptor or establish a species/isoform identity absent from the recovered Methods.

Summary: Retain broad organ development as a structural context.

Reason: The current ARBA rule contains a primate collagen-associated FunFam condition, but its target-specific activation was not reconstructed. Independent matrix and cortical ligand evidence supports contextual organ development. This broad process is secondary to the structural/ligand molecular roles and is not a universal organ-patterning activity.

Summary: Retain skeletal morphogenesis as an inherited structural context.

Reason: The exact supporting mouse collagen identities and current term records are preserved. The historical PAINT node/MSA was not reconstructed, so no claim about its evolutionary placement is added. Collagen-III-dependent cartilage/meniscus matrix structure and the broader skeletal context support retaining the developmental assertion without inventing a separate patterning signal.

Summary: Retain receptor-mediated inhibition of an immune-response context.

Reason: Actual original LAIR1 assays demonstrate collagen-dependent receptor binding and ITIM-dependent inhibition of degranulation. Collagen performs the ligand step; it is not assigned an intracellular inhibitory enzyme activity. This secondary immune process is represented alongside the core receptor-ligand function, not as a third core.

Summary: Retain postnatal cartilage development through matrix structure.

Reason: The exact mouse source examines two-week and two-month cartilage/meniscus and directly links Col3a1 reduction to abnormal fibril architecture and mechanics. Structural matrix contribution during maturation supports this contextual process without equating it with chondrocyte differentiation or a new cartilage signal.

Summary: Neither the Tsk2 growth observation nor the additional electronic premise resolves normal endochondral morphogenesis.

Reason: The current mouse source includes an IMP record from PMID:17960558 and a separate ARBA-derived IEA. The complete original abstract reports shortened Tsk2 femurs; full Methods remain unrecovered. Later full PMID:25330296 establishes the Col3a1 C33S gain-of-function allele but is not an assay of conversion of a cartilaginous anlage into bone, as required by GO:0060350. The separate electronic premise is not reduced to that one phenotype. Retain UNDECIDED because the exact normal structural contribution is unresolved, not because collagen cannot help organize developing bone or because a smaller femur disproves morphogenesis.

Summary: Retain structural organization of aortic smooth-muscle tissue.

Reason: Actual mouse Col3a1-deficient aortic ultrastructure shows abnormal relationships among collagen fibrils, elastic lamellae and smooth-muscle layers. This supports matrix-mediated tissue organization. Unchanged measured smooth-muscle contractility does not negate the structural role; no cell-fate or intrinsic contractile function is attributed to collagen.

Summary: Retain the pial-basement-membrane organization context.

Reason: Original mouse Col3a1-loss experiments show disruption of the pial membrane with neuronal overmigration and cortical dyslamination. Collagen III supplies matrix/ligand support to this structure. This process assertion does not by itself locate every collagen-III molecule within a canonical basement-membrane collagen-IV network or establish a human isoform-specific mechanism.

Summary: Retain collagen-dependent supramolecular fiber organization.

Reason: The complete original human fibroblast abstract describes disorganized collagen and fibronectin ECM in COL3A1-mutant cells. Independent human dermal and mouse fibril measurements corroborate structural participation in fiber organization. Preserve the broader source term without a redundant new process or an unobserved catalytic mechanism.

Summary: The angiotensin-response allocation remains unresolved beyond vascular susceptibility.

Reason: The current mouse IMP source reports increased dissection/rupture under angiotensin-II challenge in Col3a1-haploinsufficient mice, including a norepinephrine blood-pressure comparison. The measured susceptibility and collagen-dependent vessel integrity are real. The available original abstract does not resolve a response step performed by collagen after the stimulus, rather than a pre-existing structural vulnerability; full source text was not recovered.

Summary: Retain inhibition of neuronal migration through the ligand step.

Reason: Original collagen/GPR56 experiments measure inhibited mouse neuronal migration with receptor competition and signaling controls. This establishes performed ligand work and supports the original IEA inference. It does not establish directional chemorepulsion, a direct human neuronal assay or activity of the short human product.

Summary: Retain inhibition of neuronal migration through the ligand step.

Reason: Original collagen/GPR56 experiments measure inhibited mouse neuronal migration with receptor competition and signaling controls. This establishes performed ligand work and supports the original ISS inference. It does not establish directional chemorepulsion, a direct human neuronal assay or activity of the short human product.