F13A1

Summary: Retain the inherited protein cross-linking enzyme activity.

Reason: The actual PTHR11590 tree places human P00488 at PTN002480698 below the asserted PTN000904289 node; its aligned sequence matches the displayed 732-residue protein exactly. The corresponding IBD asserts GO:0003810, with no relevant NOT on the target path. Human F13A1 expressed from sequence-verified NM_000129.3 cDNA in COS-1 cells catalyzes amine incorporation and fibrin-chain cross-linking in PMID:27363989. This agrees with the current term chemistry. The target in its own WITH list is legitimate experimental descendant evidence. The tree service’s separate negative generic GO-slim display is not the authoritative IBD assertion and is not evidence that F13A1 lacks catalysis.

Summary: Retain the directly assayed human transglutaminase activity.

Reason: The complete PMID:27363989 article and supplement reproduced in the author thesis describe sequence-checked human F13A1 cDNA expressed in COS-1 cells. Activated wild-type preparations incorporate a primary amine into fibrinogen and incorporate alpha2-antiplasmin; calcium-dependent activity and the fibrin cross-linking assays establish the protein-glutamine gamma-glutamyltransferase reaction. COS-1 is the expression host, not the species of the insert. The 100 mM calcium experiment without thrombin is a separate in vitro activation condition, not a normal plasma calcium requirement. No activity equivalence is inferred for every patient variant or external transcript model.

Summary: Retain the convergent transglutaminase activity mappings.

Reason: Current ARBA00088940 condition sets 0 and 1 match the captured human target FunFam and taxonomy fields. All three cited InterPro entries map to GO:0003810 and occur on the actual target; RHEA:54816 and EC 2.3.2.13 specify protein glutamine/lysine isopeptide cross-linking with ammonium release. The mouse Q8BH61 identity and its experimental annotation sources are verified, although only the original abstracts of PMID:12529747 and PMID:18224415 were accessible. Independent human F13A1 amine-incorporation and fibrin-cross-linking assays in PMID:27363989 establish the target activity without relying on an uninspected mouse Methods section.

Summary: Retain the established activity while distinguishing structural evidence.

Reason: PMID:2877456 identifies and sequences human placental factor XIII A and describes human plasma material, but explicitly states that the structural study made no activity measurements. The TAS statement is biologically correct and is independently supported by the executed human wild-type amine-incorporation and fibrin-cross-linking experiments in PMID:27363989. Retention therefore does not relabel the 1986 sequence analysis as a catalytic assay or equate its historical residue numbering with the current 732-residue sequence.

Summary: Retain the reported O15273 interaction as non-core.

Reason: The actual Haenig NeuroNet source and current IntAct records report F13A1/P00488 with O15273 (TCAP) in two-hybrid pooling, array and validated two-hybrid contexts. The supplement identifies F13A1 bait clone CCSB_8309; the Methods describe human cDNA or fragment fusion constructs, exclusion of autoactivating baits, three yeast reporters and repeated pairwise tests. This supports the reported physical association, not a demonstrated physiological transglutaminase substrate, covalent cross-link, adaptor activity or transfer of the partner's function. Specific functional binding alternatives were considered, but these experiments do not establish the coordinated molecular role required for such a replacement. Study-wide mammalian validation is not claimed for this pair. Current parent-record identity does not verify the historical insert sequence, and a partner suffix is not a F13A1 isoform annotation.

Summary: Retain the reported O60636 interaction as non-core.

Reason: The actual Haenig NeuroNet source and current IntAct records report F13A1/P00488 with O60636 (TSPAN2) in two-hybrid pooling, array and validated two-hybrid contexts. The supplement identifies F13A1 bait clone CCSB_8309; the Methods describe human cDNA or fragment fusion constructs, exclusion of autoactivating baits, three yeast reporters and repeated pairwise tests. This supports the reported physical association, not a demonstrated physiological transglutaminase substrate, covalent cross-link, adaptor activity or transfer of the partner's function. Specific functional binding alternatives were considered, but these experiments do not establish the coordinated molecular role required for such a replacement. Study-wide mammalian validation is not claimed for this pair. Current parent-record identity does not verify the historical insert sequence, and a partner suffix is not a F13A1 isoform annotation.

Summary: Retain the reported O60830 interaction as non-core.

Reason: The actual Haenig NeuroNet source and current IntAct records report F13A1/P00488 with O60830 (TIMM17B) in two-hybrid pooling, array and validated two-hybrid contexts. The supplement identifies F13A1 bait clone CCSB_8309; the Methods describe human cDNA or fragment fusion constructs, exclusion of autoactivating baits, three yeast reporters and repeated pairwise tests. This supports the reported physical association, not a demonstrated physiological transglutaminase substrate, covalent cross-link, adaptor activity or transfer of the partner's function. Specific functional binding alternatives were considered, but these experiments do not establish the coordinated molecular role required for such a replacement. Study-wide mammalian validation is not claimed for this pair. Current parent-record identity does not verify the historical insert sequence, and a partner suffix is not a F13A1 isoform annotation.

Summary: Retain the reported P19388 interaction as non-core.

Reason: The actual Haenig NeuroNet source and current IntAct records report F13A1/P00488 with P19388 (POLR2E) in two-hybrid pooling, array and validated two-hybrid contexts. The supplement identifies F13A1 bait clone CCSB_8309; the Methods describe human cDNA or fragment fusion constructs, exclusion of autoactivating baits, three yeast reporters and repeated pairwise tests. This supports the reported physical association, not a demonstrated physiological transglutaminase substrate, covalent cross-link, adaptor activity or transfer of the partner's function. Specific functional binding alternatives were considered, but these experiments do not establish the coordinated molecular role required for such a replacement. Study-wide mammalian validation is not claimed for this pair. Current parent-record identity does not verify the historical insert sequence, and a partner suffix is not a F13A1 isoform annotation.

Summary: Retain the reported Q14005-2 interaction as non-core.

Reason: The actual Haenig NeuroNet source and current IntAct records report F13A1/P00488 with Q14005-2 (IL16) in two-hybrid pooling, array and validated two-hybrid contexts. The supplement identifies F13A1 bait clone CCSB_8309; the Methods describe human cDNA or fragment fusion constructs, exclusion of autoactivating baits, three yeast reporters and repeated pairwise tests. This supports the reported physical association, not a demonstrated physiological transglutaminase substrate, covalent cross-link, adaptor activity or transfer of the partner's function. Specific functional binding alternatives were considered, but these experiments do not establish the coordinated molecular role required for such a replacement. Study-wide mammalian validation is not claimed for this pair. Current parent-record identity does not verify the historical insert sequence, and a partner suffix is not a F13A1 isoform annotation.

Summary: Retain the reported Q2NKQ1-4 interaction as non-core.

Reason: The actual Haenig NeuroNet source and current IntAct records report F13A1/P00488 with Q2NKQ1-4 (SGSM1) in two-hybrid pooling, array and validated two-hybrid contexts. The supplement identifies F13A1 bait clone CCSB_8309; the Methods describe human cDNA or fragment fusion constructs, exclusion of autoactivating baits, three yeast reporters and repeated pairwise tests. This supports the reported physical association, not a demonstrated physiological transglutaminase substrate, covalent cross-link, adaptor activity or transfer of the partner's function. Specific functional binding alternatives were considered, but these experiments do not establish the coordinated molecular role required for such a replacement. Study-wide mammalian validation is not claimed for this pair. Current parent-record identity does not verify the historical insert sequence, and a partner suffix is not a F13A1 isoform annotation.

Summary: Retain the reported Q6DKI2 interaction as non-core.

Reason: The actual Haenig NeuroNet source and current IntAct records report F13A1/P00488 with Q6DKI2 (LGALS9C) in two-hybrid pooling, array and validated two-hybrid contexts. The supplement identifies F13A1 bait clone CCSB_8309; the Methods describe human cDNA or fragment fusion constructs, exclusion of autoactivating baits, three yeast reporters and repeated pairwise tests. This supports the reported physical association, not a demonstrated physiological transglutaminase substrate, covalent cross-link, adaptor activity or transfer of the partner's function. Specific functional binding alternatives were considered, but these experiments do not establish the coordinated molecular role required for such a replacement. Study-wide mammalian validation is not claimed for this pair. Current parent-record identity does not verify the historical insert sequence, and a partner suffix is not a F13A1 isoform annotation.

Summary: Retain the reported Q6NXG1 interaction as non-core.

Reason: The actual Haenig NeuroNet source and current IntAct records report F13A1/P00488 with Q6NXG1 (ESRP1) in two-hybrid pooling, array and validated two-hybrid contexts. The supplement identifies F13A1 bait clone CCSB_8309; the Methods describe human cDNA or fragment fusion constructs, exclusion of autoactivating baits, three yeast reporters and repeated pairwise tests. This supports the reported physical association, not a demonstrated physiological transglutaminase substrate, covalent cross-link, adaptor activity or transfer of the partner's function. Specific functional binding alternatives were considered, but these experiments do not establish the coordinated molecular role required for such a replacement. Study-wide mammalian validation is not claimed for this pair. Current parent-record identity does not verify the historical insert sequence, and a partner suffix is not a F13A1 isoform annotation.

Summary: Retain the reported Q6NXG1-3 interaction as non-core.

Reason: The actual Haenig NeuroNet source and current IntAct records report F13A1/P00488 with Q6NXG1-3 (ESRP1) in two-hybrid pooling, array and validated two-hybrid contexts. The supplement identifies F13A1 bait clone CCSB_8309; the Methods describe human cDNA or fragment fusion constructs, exclusion of autoactivating baits, three yeast reporters and repeated pairwise tests. This supports the reported physical association, not a demonstrated physiological transglutaminase substrate, covalent cross-link, adaptor activity or transfer of the partner's function. Specific functional binding alternatives were considered, but these experiments do not establish the coordinated molecular role required for such a replacement. Study-wide mammalian validation is not claimed for this pair. Current parent-record identity does not verify the historical insert sequence, and a partner suffix is not a F13A1 isoform annotation.

Summary: Retain the reported Q6P4D5-2 interaction as non-core.

Reason: The actual Haenig NeuroNet source and current IntAct records report F13A1/P00488 with Q6P4D5-2 (PABIR3) in two-hybrid pooling, array and validated two-hybrid contexts. The supplement identifies F13A1 bait clone CCSB_8309; the Methods describe human cDNA or fragment fusion constructs, exclusion of autoactivating baits, three yeast reporters and repeated pairwise tests. This supports the reported physical association, not a demonstrated physiological transglutaminase substrate, covalent cross-link, adaptor activity or transfer of the partner's function. Specific functional binding alternatives were considered, but these experiments do not establish the coordinated molecular role required for such a replacement. Study-wide mammalian validation is not claimed for this pair. Current parent-record identity does not verify the historical insert sequence, and a partner suffix is not a F13A1 isoform annotation.

Summary: Retain the reported Q6ZMI0-5 interaction as non-core.

Reason: The actual Haenig NeuroNet source and current IntAct records report F13A1/P00488 with Q6ZMI0-5 (PPP1R21) in two-hybrid pooling, array and validated two-hybrid contexts. The supplement identifies F13A1 bait clone CCSB_8309; the Methods describe human cDNA or fragment fusion constructs, exclusion of autoactivating baits, three yeast reporters and repeated pairwise tests. This supports the reported physical association, not a demonstrated physiological transglutaminase substrate, covalent cross-link, adaptor activity or transfer of the partner's function. Specific functional binding alternatives were considered, but these experiments do not establish the coordinated molecular role required for such a replacement. Study-wide mammalian validation is not claimed for this pair. Current parent-record identity does not verify the historical insert sequence, and a partner suffix is not a F13A1 isoform annotation.

Summary: Retain the reported Q7Z783 interaction as non-core.

Reason: The actual Haenig NeuroNet source and current IntAct records report F13A1/P00488 with Q7Z783 in two-hybrid pooling, array and validated two-hybrid contexts. The supplement identifies F13A1 bait clone CCSB_8309; the Methods describe human cDNA or fragment fusion constructs, exclusion of autoactivating baits, three yeast reporters and repeated pairwise tests. This supports the reported physical association, not a demonstrated physiological transglutaminase substrate, covalent cross-link, adaptor activity or transfer of the partner's function. Specific functional binding alternatives were considered, but these experiments do not establish the coordinated molecular role required for such a replacement. Study-wide mammalian validation is not claimed for this pair. Current parent-record identity does not verify the historical insert sequence, and a partner suffix is not a F13A1 isoform annotation. Q7Z783 remains the literal human accession: its current unreviewed record lacks a GeneID, so no current-gene table join is asserted.

Summary: Retain the reported Q8IUR5-4 interaction as non-core.

Reason: The actual Haenig NeuroNet source and current IntAct records report F13A1/P00488 with Q8IUR5-4 (TMTC1) in two-hybrid pooling, array and validated two-hybrid contexts. The supplement identifies F13A1 bait clone CCSB_8309; the Methods describe human cDNA or fragment fusion constructs, exclusion of autoactivating baits, three yeast reporters and repeated pairwise tests. This supports the reported physical association, not a demonstrated physiological transglutaminase substrate, covalent cross-link, adaptor activity or transfer of the partner's function. Specific functional binding alternatives were considered, but these experiments do not establish the coordinated molecular role required for such a replacement. Study-wide mammalian validation is not claimed for this pair. Current parent-record identity does not verify the historical insert sequence, and a partner suffix is not a F13A1 isoform annotation.

Summary: Retain the reported Q8WXK3-2 interaction as non-core.

Reason: The actual Haenig NeuroNet source and current IntAct records report F13A1/P00488 with Q8WXK3-2 (ASB13) in two-hybrid pooling, array and validated two-hybrid contexts. The supplement identifies F13A1 bait clone CCSB_8309; the Methods describe human cDNA or fragment fusion constructs, exclusion of autoactivating baits, three yeast reporters and repeated pairwise tests. This supports the reported physical association, not a demonstrated physiological transglutaminase substrate, covalent cross-link, adaptor activity or transfer of the partner's function. Specific functional binding alternatives were considered, but these experiments do not establish the coordinated molecular role required for such a replacement. Study-wide mammalian validation is not claimed for this pair. Current parent-record identity does not verify the historical insert sequence, and a partner suffix is not a F13A1 isoform annotation.

Summary: Retain the reported Q8WZ73-3 interaction as non-core.

Reason: The actual Haenig NeuroNet source and current IntAct records report F13A1/P00488 with Q8WZ73-3 (RFFL) in two-hybrid pooling, array and validated two-hybrid contexts. The supplement identifies F13A1 bait clone CCSB_8309; the Methods describe human cDNA or fragment fusion constructs, exclusion of autoactivating baits, three yeast reporters and repeated pairwise tests. This supports the reported physical association, not a demonstrated physiological transglutaminase substrate, covalent cross-link, adaptor activity or transfer of the partner's function. Specific functional binding alternatives were considered, but these experiments do not establish the coordinated molecular role required for such a replacement. Study-wide mammalian validation is not claimed for this pair. Current parent-record identity does not verify the historical insert sequence, and a partner suffix is not a F13A1 isoform annotation.

Summary: Retain the reported Q96EF9 interaction as non-core.

Reason: The actual Haenig NeuroNet source and current IntAct records report F13A1/P00488 with Q96EF9 (ZHX1-C8orf76) in two-hybrid pooling, array and validated two-hybrid contexts. The supplement identifies F13A1 bait clone CCSB_8309; the Methods describe human cDNA or fragment fusion constructs, exclusion of autoactivating baits, three yeast reporters and repeated pairwise tests. This supports the reported physical association, not a demonstrated physiological transglutaminase substrate, covalent cross-link, adaptor activity or transfer of the partner's function. Specific functional binding alternatives were considered, but these experiments do not establish the coordinated molecular role required for such a replacement. Study-wide mammalian validation is not claimed for this pair. Current parent-record identity does not verify the historical insert sequence, and a partner suffix is not a F13A1 isoform annotation. The original gene-level table names C8orf76/84933, whereas the current Q96EF9 record is ZHX1-C8orf76/100533106; that naming/version boundary is retained rather than equating the constructs.

Summary: Retain the reported Q96GL9 interaction as non-core.

Reason: The actual Haenig NeuroNet source and current IntAct records report F13A1/P00488 with Q96GL9 (FAM163A) in two-hybrid pooling, array and validated two-hybrid contexts. The supplement identifies F13A1 bait clone CCSB_8309; the Methods describe human cDNA or fragment fusion constructs, exclusion of autoactivating baits, three yeast reporters and repeated pairwise tests. This supports the reported physical association, not a demonstrated physiological transglutaminase substrate, covalent cross-link, adaptor activity or transfer of the partner's function. Specific functional binding alternatives were considered, but these experiments do not establish the coordinated molecular role required for such a replacement. Study-wide mammalian validation is not claimed for this pair. Current parent-record identity does not verify the historical insert sequence, and a partner suffix is not a F13A1 isoform annotation.

Summary: Retain the reported Q96MX0 interaction as non-core.

Reason: The actual Haenig NeuroNet source and current IntAct records report F13A1/P00488 with Q96MX0 (CMTM3) in two-hybrid pooling, array and validated two-hybrid contexts. The supplement identifies F13A1 bait clone CCSB_8309; the Methods describe human cDNA or fragment fusion constructs, exclusion of autoactivating baits, three yeast reporters and repeated pairwise tests. This supports the reported physical association, not a demonstrated physiological transglutaminase substrate, covalent cross-link, adaptor activity or transfer of the partner's function. Specific functional binding alternatives were considered, but these experiments do not establish the coordinated molecular role required for such a replacement. Study-wide mammalian validation is not claimed for this pair. Current parent-record identity does not verify the historical insert sequence, and a partner suffix is not a F13A1 isoform annotation.

Summary: Retain the reported Q99944 interaction as non-core.

Reason: The actual Haenig NeuroNet source and current IntAct records report F13A1/P00488 with Q99944 (EGFL8) in two-hybrid pooling, array and validated two-hybrid contexts. The supplement identifies F13A1 bait clone CCSB_8309; the Methods describe human cDNA or fragment fusion constructs, exclusion of autoactivating baits, three yeast reporters and repeated pairwise tests. This supports the reported physical association, not a demonstrated physiological transglutaminase substrate, covalent cross-link, adaptor activity or transfer of the partner's function. Specific functional binding alternatives were considered, but these experiments do not establish the coordinated molecular role required for such a replacement. Study-wide mammalian validation is not claimed for this pair. Current parent-record identity does not verify the historical insert sequence, and a partner suffix is not a F13A1 isoform annotation.

Summary: Retain the reported Q9BXL8 interaction as non-core.

Reason: The actual Haenig NeuroNet source and current IntAct records report F13A1/P00488 with Q9BXL8 (CDCA4) in two-hybrid pooling, array and validated two-hybrid contexts. The supplement identifies F13A1 bait clone CCSB_8309; the Methods describe human cDNA or fragment fusion constructs, exclusion of autoactivating baits, three yeast reporters and repeated pairwise tests. This supports the reported physical association, not a demonstrated physiological transglutaminase substrate, covalent cross-link, adaptor activity or transfer of the partner's function. Specific functional binding alternatives were considered, but these experiments do not establish the coordinated molecular role required for such a replacement. Study-wide mammalian validation is not claimed for this pair. Current parent-record identity does not verify the historical insert sequence, and a partner suffix is not a F13A1 isoform annotation.

Summary: Retain the reported Q9NZC7-5 interaction as non-core.

Reason: The actual Haenig NeuroNet source and current IntAct records report F13A1/P00488 with Q9NZC7-5 (WWOX) in two-hybrid pooling, array and validated two-hybrid contexts. The supplement identifies F13A1 bait clone CCSB_8309; the Methods describe human cDNA or fragment fusion constructs, exclusion of autoactivating baits, three yeast reporters and repeated pairwise tests. This supports the reported physical association, not a demonstrated physiological transglutaminase substrate, covalent cross-link, adaptor activity or transfer of the partner's function. Specific functional binding alternatives were considered, but these experiments do not establish the coordinated molecular role required for such a replacement. Study-wide mammalian validation is not claimed for this pair. Current parent-record identity does not verify the historical insert sequence, and a partner suffix is not a F13A1 isoform annotation.

Summary: Retain the reported Q9Y2M5 interaction as non-core.

Reason: The actual Haenig NeuroNet source and current IntAct records report F13A1/P00488 with Q9Y2M5 (KLHL20) in two-hybrid pooling, array and validated two-hybrid contexts. The supplement identifies F13A1 bait clone CCSB_8309; the Methods describe human cDNA or fragment fusion constructs, exclusion of autoactivating baits, three yeast reporters and repeated pairwise tests. This supports the reported physical association, not a demonstrated physiological transglutaminase substrate, covalent cross-link, adaptor activity or transfer of the partner's function. Specific functional binding alternatives were considered, but these experiments do not establish the coordinated molecular role required for such a replacement. Study-wide mammalian validation is not claimed for this pair. Current parent-record identity does not verify the historical insert sequence, and a partner suffix is not a F13A1 isoform annotation.

Summary: Retain the extracellular plasma pool.

Reason: The accessible primary summary of PMID:4405643 explicitly reports purification and comparison of human plasma and platelet factor XIII, including plasma A2B2. The complete original Methods were not recovered, so their exact purification and contamination controls are not newly claimed. Independent actual human plasma material in PMID:2877456 corroborates the extracellular pool. This broad location is retained without inferring a conventional signal peptide, an active secretion mechanism or that all cellular F13A1 is extracellular.

Summary: Retain the extracellular plasma and clot-associated pool.

Reason: The actual UniProt SL-0243 mapping leads to extracellular region, and F13A1 participates in extracellular plasma factor XIII and clot cross-linking. This location is supported independently of the unconventional release mechanism or the abundant intracellular platelet pool. It does not imply a classical signal peptide, that all cellular F13A1 is secreted, or that processing chains are natural isoforms. Human F13A1-dependent cross-linking of fibrin in deficient plasma is directly assayed in PMID:27363989.

Summary: Retain the extracellular pool independently of a secretion mechanism.

Reason: PMID:2877457 is the actual human placental cDNA study and describes circulating plasma factor XIII in its biological context. Its cloning and RNA experiments do not directly measure secretion, and the paper discusses the lack of an obvious signal peptide. Human plasma protein studies independently support extracellular residence. Retain the TAS location without converting a plasma observation into evidence for a particular export route or treating placental RNA as a plasma localization assay.

Summary: Retain extracellular F13A1 during thrombin-dependent activation.

Reason: R-HSA-140599 explicitly contains human P00488 in the plasma A2B2 substrate, cleaved tetramer and released activation peptide, all in extracellular region. Thrombin is the catalyst; F13A1 is the cleavage substrate in this event, so the location does not transfer protease activity to it. Actual entity coordinates are 1–732, 39–732 and 1–38, while some display names retain historical 731-based numbering. The coordinate/display discrepancy is preserved rather than silently repaired.

Summary: Retain extracellular F13A1 during calcium-dependent complex activation.

Reason: R-HSA-140847 resolves the cleaved plasma tetramer to an extracellular calcium-associated F13A1 product and released B subunits. The target chain is explicitly P00488 residues 39–732. This is a modeled complex-dissociation and activation step, not proof that F13A1 catalyzes its own proteolytic activation or that the model establishes every active-state stoichiometry. The extracellular location of the actual target is supported.

Summary: Retain extracellular activated F13A1 as the fibrin cross-linking catalyst.

Reason: R-HSA-140851 catalystActivity 140850 resolves to factor XIIIa entity 140849 containing human P00488 residues 39–732 in extracellular region. F13A1 performs the cross-linking reaction rather than merely being a substrate needed for clotting. The independent human expression and fibrin-chain cross-linking experiments in PMID:27363989 corroborate this role. Reactome’s historical display numbering is kept distinct from the actual coordinate fields.

Summary: Retain the extracellular product location in platelet-content release.

Reason: The complete R-HSA-481007 participant graph connects alpha-granule F13A1 entity 349759 to extracellular P00488 entity 140600 in the released-content set. This supports the imported extracellular location without making exocytosis a performed F13A1 process. The minor reported plasma-derived granule pool must be distinguished from the much larger cytoplasmic platelet pool: PMID:25331118 and PMID:39116293 directly show that platelet F13A1 retention, surface exposure and release depend on preparation and stimulation. The source is not used to assert that most platelet F13A1 follows alpha-granule secretion.

Summary: Retain the explicit extracellular F13A1 product location.

Reason: R-HSA-6785895 places human P00488 entity 140600 in the extracellular protein output of an IL4/IL13-responsive expression event. F13A1 is a product of the regulated expression model, not the cytokine, transcriptional regulator or enzyme driving protein synthesis. The explicit target compartment agrees with its independently established extracellular clot function; the source does not establish a universal cellular secretion pathway or justify a new inflammatory-regulation annotation.

Summary: Retain extracellular F13A1 catalysis on platelet-bound fibrin.

Reason: R-HSA-9921091 uses the same extracellular factor XIIIa catalyst 140849/140850 to cross-link fibrin associated with platelet integrin alphaIIb beta3. The actual target chain is P00488 residues 39–732; the membrane-associated fibrin substrate and integrin do not make F13A1 an integral membrane protein. Retain the explicitly modeled extracellular catalyst location without inferring a new membrane topology or treating all platelet-associated pools as equivalent.

Summary: Specify the soluble cytosolic platelet pool.

Reason: SL-0086 correctly maps to broad cytoplasm. In the actual human platelet Methods of PMID:25331118, sonicated lysate is cleared at 12,000g and then separated at 100,000g; the soluble supernatant designated cytoplasmic fraction contains F13A1 antigen and most measured transglutaminase activity, with a distinct membrane pellet. That experimentally resolved soluble pool fits current GO:0005829 cytosol more specifically than cytoplasm. PMID:39116293 independently supports substantial intracellular retention after dual-agonist stimulation, while preserving separate surface and released pools. This refinement does not deny the minor granule pool or assign all F13A1 to cytosol.

Proposed replacements: cytosol

Summary: Specify the performed fibrin-clot formation step.

Reason: The central coagulation contribution of F13A1 is enzymatic stabilization of fibrin by cross-linking, captured by current GO:0072378, which is part of blood coagulation. PMID:10801785 experimentally measures bovine thrombin cleavage of synthetic human factor XIII activation peptides; F13A1 is the substrate in that reaction. Independent human F13A1 experiments in PMID:27363989 measure gamma-chain dimers, alpha-chain multimers and clot architecture after activation and directly support the downstream fibrin-clot formation step performed by F13A1. The refinement does not give F13A1 thrombin activity or make the V34L cardioprotection hypothesis an established mechanism.

Proposed replacements: blood coagulation, fibrin clot formation

Summary: Retain the executed peptide cross-linking reaction.

Reason: Human wild-type F13A1 preparations in PMID:27363989 generate fibrin gamma-chain dimers and alpha-chain multimers in factor-XIII-deficient plasma supplemented with thrombin and calcium; normal plasma and recombinant FXIII-A controls are described in the supplement. Amine and alpha2-antiplasmin incorporation provide complementary reaction readouts. F13A1 performs the cross-linking, while thrombin performs its proteolytic activation. The substrate and activation controls do not establish identical behavior for every variant or natural transcript.

Summary: Retain colon ECM detection as a contextual location.

Reason: The actual human tissue study PMID:25037231 uses sequential cellular-protein depletion followed by LC-MS/MS of the insoluble ECM-enriched fraction. Additional file 2 identifies F13A1/P00488 in normal colon and colon tumors, including multiple unique peptides. The raw GOA colon extension is retained in the source-context record. This positive tissue-associated location is useful but does not by itself establish active transglutaminase, a particular cellular source, or a colon-specific mechanism; it is retained as non-core relative to the established clot-stabilizing chemistry.

Summary: Retain bone-marrow ECM detection as a contextual location.

Reason: PMID:28344315 profiles ECM-enriched human bone marrow from healthy donors and multiple-myeloma groups after sequential cellular-protein depletion. The actual Table S3 identifies human F13A1/P00488 in the complete dataset and matrisome subsets, with 18 unique peptides in the complete dataset. Retain this positive fraction-localization evidence and its raw bone-marrow extension as non-core. Neither the mass-spectrometry identification nor a disease-group comparison establishes enzyme activation, a particular producer cell or a causal myeloma function.

Summary: Retain omental ECM detection as a contextual location.

Reason: PMID:28675934 studies human normal omentum and ovarian-cancer omental metastases using sequential depletion of cellular components, followed by crude ECM or urea-soluble ECM digestion and LC-MS/MS. Its actual Supplementary Table S1 lists P00488/F13A1 in the omental samples, including the fractionated normal-omentum datasets. Here crude means an ECM-enriched preparation, not untreated whole tissue. Retain the positive location and raw omentum extension as non-core; the data do not assign activation state, cross-linking substrates in those tissues or a cancer mechanism.

Summary: Retain the reported minor alpha-granule lumen pool as contextual.

Reason: R-HSA-481007 explicitly includes P00488 residues 1–732 in alpha-granule lumen entity 349759. The accessible primary abstract of PMID:8360253 reports density-fractionated human alpha granules containing A and B subunits and thrombin-inducible factor XIII activity; its full original Methods were unavailable. The independently read PMID:25331118 and PMID:39116293 distinguish a small plasma-derived A2B2 granule pool from the predominant intracellular platelet A-subunit pool. Retain this positive source-specific location as non-core rather than removing it because cytoplasmic F13A1 is more abundant. The particular uptake route, fraction purity and contribution of this minor pool are not newly established here.

Summary: Retain ancestral participation in fibrin-clot formation.

Reason: The actual human target descends from PTN002612883, whose positive IBD for GO:0072378 is grounded in mouse MGI:1921395 and human P00488. No relevant target-path IBD NOT was found. Human F13A1 expressed from the 732-residue cDNA directly promotes gamma-chain dimers and alpha-chain multimers in fibrin cross-linking assays with deficient plasma (PMID:27363989). Thus the product performs a clot-stabilizing enzymatic step, not merely a necessary substrate role. Mouse-source abstracts are consistent but their unread full Methods are not substituted for this target-specific evidence.

Summary: Retain the catalytic contribution to fibrin-clot formation.

Reason: The full human construct and deficient-plasma experiments in PMID:27363989 directly test F13A1-dependent fibrin cross-linking and resulting clot structure after activation. The product performs the stabilization chemistry rather than merely being consumed by coagulation. The two raw GOA records collapse to this one context only because annotation identifiers, dates and assigners differ; their biological source fields are preserved. This does not assert that F13A1 performs upstream thrombin generation or its own proteolytic activation.

Summary: Retain the rule-supported fibrin-clot formation process.

Reason: Current ARBA00090798 condition set 0 matches the human target’s three FunFam features and vertebrate lineage. The official Q8BH61/ENSMUSP00000048667 donor bridge and mouse experimental sources are verified; their accessible abstracts report loss of cross-linking in deficient animals, with full original access limited. Independently, the human F13A1 amine-incorporation and fibrin-cross-linking experiments in PMID:27363989 demonstrate the performed enzymatic contribution to clot formation. This does not confuse F13A1’s own thrombin-dependent maturation with the cross-linking work it performs.

Summary: Leave the target-specific microparticle assignment unresolved.

Reason: PMID:22516433 reports a proteomic survey of microvesicles from healthy human plasma, but the complete operative Methods and the F13A1-specific target table were not recovered. The previously saved attachment resolves to an error response rather than a usable workbook. The abstract-level survey cannot verify the accession, peptide assignment, vesicle isolation specificity or processed state for this row. Preserve the experimental source assertion as UNDECIDED rather than rejecting it from an unavailable table, and do not infer microparticle localization merely from soluble plasma F13A1.

Summary: Retain membership in a catalytic transferase complex.

Reason: PMID:27363989 describes cellular FXIII-A2 and plasma FXIII-A2B2, with catalytic A subunits and carrier B subunits, and tests expressed human A-subunit function in clotting preparations. The older human protein study independently distinguishes plasma and platelet complexes. The broad current transferase-complex term captures participation of the catalytic A subunit without assigning catalysis to F13B or inferring that every assay preparation retained an identical oligomeric state. No unsupported more specific complex identifier is substituted.

Summary: Human factor XIII A subunits directly bind calcium in the experimentally determined A2 structure.

Reason: Calcium binding is established independently of calcium-dependent catalytic activity: the primary deposited X-ray structure 1GGU, associated with PMID:9988734, contains two modeled calcium ions in recombinant human A2 produced in yeast. Its deposited 731-residue polymer sequence exactly matches current P00488 residues 2–732. The publication abstract identifies a calcium site in each monomer, and UniProt FT BINDING and CC COFACTOR independently curate the same paper and current-sequence side-chain positions 437, 439, 486 and 491. These are binding observations on the A subunit, not merely a requirement for calcium in an activation mixture. The current unrestricted GO:0005509 definition fits this molecular function and is not an ancestor or descendant of the retained catalytic activity. Exact-term comparison found IDA annotations on human TGM2 and TGM3, with no result for TGM1 or F13A1 in that query; this argues against systematic exclusion of calcium binding from transglutaminases but is not the experimental evidence for F13A1. The whole article Methods and Results remain inaccessible; the assertion relies on the inspected primary structure deposition, abstract and curated UniProt record. The observed crystal-state occupancy is not generalized to all activated assemblies, additional calcium sites or untested transcript products.