EFEMP1

Summary: EFEMP1 associates with EGFR rather than acting as the receptor.

Reason: In human pancreatic carcinoma cells, EFEMP1 co-immunoprecipitated with EGFR and EGF treatment reduced the association. Purified human EFEMP1 treatment increased receptor and downstream phosphorylation. These results support EGFR binding, not an EFEMP1 transmembrane receptor or kinase activity. The association was measured in cell lysates, not as a purified binary interaction; no natural splice product was identified.

Proposed replacements: epidermal growth factor receptor binding

Summary: EGFR association is a demonstrated, context-dependent EFEMP1 activity.

Reason: The recovered original reports EFEMP1–EGFR co-immunoprecipitation with a no-antibody control, reduced association after EGF exposure, and EGFR-dependent signaling after purified human EFEMP1 treatment. This supports binding in the measured human carcinoma context. It does not establish a purified binding interface, universal normal-tissue signaling, or equivalent activity of all splice products.

Summary: Phylogenetically inferred matrix structural activity is supported by target experiments.

Reason: The current PANTHER gene assignment resolves to PTHR24039, and human EFEMP1 descends from the literal IBD node PTN000620536. The broad assertion concerns matrix structural integrity, not fibrillin microfibril polymerization. Human ARPE-19 loss and rescue experiments independently show defective collagen VI matrix organization after EFEMP1 loss. Neither the short donor list nor evidence on a descendant is a circularity defect.

Summary: Human aortic proteomics supports EFEMP1 as a matrix glycoprotein.

Reason: The original human aorta study separates salt-soluble, detergent and guanidine-extracted material. Its FBLN3_HUMAN rows report multiple unique peptides, including 38 unique peptides and 22.0% coverage in Table I. The structural classification is an RCA interpretation of an extracellular matrix glycoprotein, consistent with the GO definition and later human matrix knockout/rescue evidence; mass spectrometry alone is not a mechanical-strength assay.

Summary: The human tissue matrisome contains an EFEMP1 protein group.

Reason: The recovered supplementary workbook lists EFEMP1 as a core-matrisome ECM glycoprotein in normal colon, colorectal tumor, liver metastasis and normal liver contexts. The accession group includes Q12805 and several splice-product identifiers, so it does not distinguish the five current products. This RCA structural classification is consistent with the term definition and independent human EFEMP1 matrix experiments.

Summary: The matrix structural classification is consistent with the curated proteomics assertion.

Reason: The original describes ECM enrichment and mass spectrometry of human breast and omental tissues and tumors, with explicit peptide filtering and matrisome classification. The EFEMP1-specific supplementary row was not recovered. The broad structural assignment is nevertheless supported by independently recovered human EFEMP1 matrix measurements and the GO definition permitting ECM glycoproteins; the original target row is not represented as independently inspected.

Summary: Calcium-binding EGF modules support this molecular property.

Reason: The cited InterPro domain/site mappings explicitly supply calcium ion binding, and the official EFEMP1 record annotates calcium-binding EGF-like modules. This is a domain-based inference, not a target-specific calcium affinity measurement, and it does not assign identical calcium handling to every splice product.

Summary: Binding to ARAF supports a verified partner-class refinement.

Reason: 3 exact source/partner record(s) were recovered from IntAct/MINT, with their assay methods and feature intervals preserved. ARAF is an established protein kinase. The recorded binding-associated regions are N-terminal regulatory segments (66–155, 2–314 and 2–350), so the result identifies binding to a kinase protein without claiming catalytic-domain recognition or EFEMP1 kinase activity.

Proposed replacements: protein kinase binding

Summary: Binding to SGTA supports a verified partner-class refinement.

Reason: 1 exact source/partner record(s) were recovered from IntAct/MINT, with their assay methods and feature intervals preserved. SGTA is a substrate-binding chaperone that binds exposed hydrophobic regions and maintains proteins in a soluble, targeting-competent state. Its verified class satisfies chaperone binding without requiring an independently demonstrated EFEMP1 folding mechanism.

Proposed replacements: protein-folding chaperone binding

Summary: Binding to ATN1 supports a verified partner-class refinement.

Reason: The three source-linked records are yeast two-hybrid physical associations with human ATN1, using sufficient-binding regions 1–1185, 372–800 and 785–1190. ATN1 is a transcriptional corepressor, supporting transcription coregulator binding. The parent-sequence checksum identifies the reference protein, not the exact experimental inserts. These assays do not establish a native EFEMP1–ATN1 transcriptional complex, nuclear localization of EFEMP1, or EFEMP1 transcriptional activity; the different records are not independent native biological replications.

Proposed replacements: transcription coregulator binding

Summary: Binding to HCV NS5A is retained as a non-core observation.

Reason: The target pair is present in the recovered original supplementary table; the current IntAct/MINT query did not return this source pair. The imported processed-chain identifier denotes NS5A, not the viral RNA polymerase or protease. Original Supplementary Table SI reports EFEMP1–NS5A in the IMAP2 screen and a positive co-affinity-purification result. No informative supported partner-class term is established; the positive interaction is retained without assigning viral catalytic functions to EFEMP1.

Summary: Binding to ECM1 supports a verified partner-class refinement.

Reason: The source-specific abstract explicitly identifies EFEMP1/fibulin-3; current IntAct/MINT query absence is not treated as evidence against binding. The source abstract explicitly reports a human foreskin-library Y2H interaction followed by in vitro and in vivo co-immunoprecipitation and human-skin colocalization. ECM1 is an extracellular matrix protein, meeting the matrix-binding definition. Full assay conditions were not recovered, and colocalization or co-immunoprecipitation is not described as a purified interface.

Proposed replacements: extracellular matrix binding

Summary: Binding to HOXA1 supports a verified partner-class refinement.

Reason: The source-linked record is a yeast two-hybrid pooling assay with human HOXA1, with a sufficient-binding feature at residues 1–110. This segment is outside the parent homeobox at 229–288; binding to a segment of a DNA-binding transcription factor supports the partner class without requiring recognition of its DNA-binding domain. The reference-sequence checksum does not authenticate the experimental insert. The study describes a selected mammalian-validation subset, but pair-specific membership is unverified; native EFEMP1–HOXA1 colocalization and EFEMP1 transcriptional activity are not established.

Proposed replacements: DNA-binding transcription factor binding

Summary: Binding to SGTA supports a verified partner-class refinement.

Reason: 3 exact source/partner record(s) were recovered from IntAct/MINT, with their assay methods and feature intervals preserved. SGTA is a substrate-binding chaperone that binds exposed hydrophobic regions and maintains proteins in a soluble, targeting-competent state. Its verified class satisfies chaperone binding without requiring an independently demonstrated EFEMP1 folding mechanism.

Proposed replacements: protein-folding chaperone binding

Summary: Binding to BAG6 supports a verified partner-class refinement.

Reason: 3 exact source/partner record(s) were recovered from IntAct/MINT, with their assay methods and feature intervals preserved. BAG6 is an ATP-independent molecular chaperone that binds hydrophobic substrates and maintains their solubility. The positive pair therefore supports chaperone binding; no client-folding mechanism for EFEMP1 is inferred.

Proposed replacements: protein-folding chaperone binding

Summary: Binding to TRAF2 supports a verified partner-class refinement.

Reason: 3 exact source/partner record(s) were recovered from IntAct/MINT, with their assay methods and feature intervals preserved. TRAF2 has experimentally supported ubiquitin-protein ligase activity. The measured association supports ligase binding without implying that EFEMP1 catalyzes ubiquitination or is a ubiquitination substrate.

Proposed replacements: ubiquitin protein ligase binding

Summary: Binding to TXNDC5 sequence Q86UY0 supports a verified partner-class refinement.

Reason: 3 exact source/partner record(s) were recovered from IntAct/MINT, with their assay methods and feature intervals preserved. The literal 360-residue Q86UY0 sequence matches its interaction-record checksum. Residues 17–360 equal reviewed TXNDC5 residues 89–432 and retain all three CGHC motifs and the two complete C-terminal thioredoxin domains, despite a distinct N terminus. Protein-disulfide-isomerase class is inferred from this conserved catalytic sequence and official annotation, not from a direct activity assay of this historical clone.

Proposed replacements: enzyme binding

Summary: Binding to RIC8A product 4 supports a verified partner-class refinement.

Reason: 3 exact source/partner record(s) were recovered from IntAct/MINT, with their assay methods and feature intervals preserved. RIC8A is an established folding chaperone for G-protein alpha subunits. The exact product-4 sequence matches the record checksum and differs from the reference by deletion of residue 209. Two records separately annotate sufficient binding region 111–530; the parent checksum does not establish a full-length experimental construct. The refinement names the partner class, not an EFEMP1 folding or G-protein signaling role.

Proposed replacements: protein-folding chaperone binding

Summary: Binding to SGTA supports a verified partner-class refinement.

Reason: 3 exact source/partner record(s) were recovered from IntAct/MINT, with their assay methods and feature intervals preserved. SGTA is a substrate-binding chaperone that binds exposed hydrophobic regions and maintains proteins in a soluble, targeting-competent state. Its verified class satisfies chaperone binding without requiring an independently demonstrated EFEMP1 folding mechanism.

Proposed replacements: protein-folding chaperone binding

Summary: Binding to MEOX2 supports a verified partner-class refinement.

Reason: Three source-linked records identify human MEOX2 in validated, array and pooling yeast two-hybrid assays. The matched parent sequence is a DNA-binding transcription factor; two records annotate an EFEMP1 residue-345 mutation-with-no-effect feature. This supports the partner-class binding refinement while preserving the recorded variant scope, without claiming all natural products behave equivalently. Repeated screening stages do not establish native EFEMP1–MEOX2 colocalization, an endogenous transcriptional complex, or transcription-factor activity of EFEMP1.

Proposed replacements: DNA-binding transcription factor binding

Summary: Binding to CRYBA4 is retained as a non-core observation.

Reason: 3 exact source/partner record(s) were recovered from IntAct/MINT, with their assay methods and feature intervals preserved. CRYBA4 is a lens structural protein, not thereby an actin-, microtubule- or intermediate-filament component. The records include positive binding and mutation-sensitive interaction features; no informative supported refinement follows from lens structural identity alone.

Summary: Binding to TRAF2 supports a verified partner-class refinement.

Reason: 3 exact source/partner record(s) were recovered from IntAct/MINT, with their assay methods and feature intervals preserved. TRAF2 has experimentally supported ubiquitin-protein ligase activity. The measured association supports ligase binding without implying that EFEMP1 catalyzes ubiquitination or is a ubiquitination substrate.

Proposed replacements: ubiquitin protein ligase binding

Summary: Binding to SGTA supports a verified partner-class refinement.

Reason: 3 exact source/partner record(s) were recovered from IntAct/MINT, with their assay methods and feature intervals preserved. SGTA is a substrate-binding chaperone that binds exposed hydrophobic regions and maintains proteins in a soluble, targeting-competent state. Its verified class satisfies chaperone binding without requiring an independently demonstrated EFEMP1 folding mechanism.

Proposed replacements: protein-folding chaperone binding

Summary: Binding to TRAF2 supports a verified partner-class refinement.

Reason: 3 exact source/partner record(s) were recovered from IntAct/MINT, with their assay methods and feature intervals preserved. TRAF2 has experimentally supported ubiquitin-protein ligase activity. The measured association supports ligase binding without implying that EFEMP1 catalyzes ubiquitination or is a ubiquitination substrate.

Proposed replacements: ubiquitin protein ligase binding

Summary: Binding to historical MEOX2 sequence Q6FHY5 supports a verified partner-class refinement.

Reason: Three source-linked records identify historical MEOX2/Q6FHY5 in validated, array and prey-pooling yeast two-hybrid assays. Its 303-residue reference sequence matches the recorded checksum and equals current MEOX2/P50222 with canonical Q81 deleted; the homeobox remains. This supports the DNA-binding transcription-factor partner class while preserving the imported identifier. The checksum authenticates the reference sequence, not the experimental insert, and the assays do not establish native human colocalization, an endogenous transcriptional complex, or transcriptional activity of EFEMP1.

Proposed replacements: DNA-binding transcription factor binding

Summary: Binding to SGTA supports a verified partner-class refinement.

Reason: 1 exact source/partner record(s) were recovered from IntAct/MINT, with their assay methods and feature intervals preserved. SGTA is a substrate-binding chaperone that binds exposed hydrophobic regions and maintains proteins in a soluble, targeting-competent state. Its verified class satisfies chaperone binding without requiring an independently demonstrated EFEMP1 folding mechanism.

Proposed replacements: protein-folding chaperone binding

Summary: Binding to BAG6 supports a verified partner-class refinement.

Reason: 1 exact source/partner record(s) were recovered from IntAct/MINT, with their assay methods and feature intervals preserved. The BAG6 record includes a variant annotated as having no effect on interaction. BAG6 is an ATP-independent molecular chaperone that binds hydrophobic substrates and maintains their solubility. The positive pair therefore supports chaperone binding; no client-folding mechanism for EFEMP1 is inferred.

Proposed replacements: protein-folding chaperone binding

Summary: Binding to CYSRT1 is retained as a non-core observation.

Reason: 5 exact source/partner record(s) were recovered from IntAct/MINT, with their assay methods and feature intervals preserved. HuRI is a binary assay in a heterologous system; it does not establish native colocalization or physiological interaction. CYSRT1 has a cornified-layer/antimicrobial role, but the inspected record does not establish a more informative binding-partner class. Positive binary binding is retained without inferring antimicrobial activity for EFEMP1.

Summary: Binding to SGTA supports a verified partner-class refinement.

Reason: 3 exact source/partner record(s) were recovered from IntAct/MINT, with their assay methods and feature intervals preserved. HuRI is a binary assay in a heterologous system; it does not establish native colocalization or physiological interaction. SGTA is a substrate-binding chaperone that binds exposed hydrophobic regions and maintains proteins in a soluble, targeting-competent state. Its verified class satisfies chaperone binding without requiring an independently demonstrated EFEMP1 folding mechanism.

Proposed replacements: protein-folding chaperone binding

Summary: Binding to BAG4 is retained as a non-core observation.

Reason: 3 exact source/partner record(s) were recovered from IntAct/MINT, with their assay methods and feature intervals preserved. HuRI is a binary assay in a heterologous system; it does not establish native colocalization or physiological interaction. BAG4 is an HSP70 co-chaperone/nucleotide-exchange regulator. That role alone does not demonstrate the direct nascent/unfolded-substrate-binding class specified by GO:0051087. The pair remains positive; no EFEMP1 client mechanism is demanded or inferred.

Summary: Binding to SLPI is retained as a non-core observation.

Reason: 3 exact source/partner record(s) were recovered from IntAct/MINT, with their assay methods and feature intervals preserved. HuRI is a binary assay in a heterologous system; it does not establish native colocalization or physiological interaction. SLPI is a protease inhibitor rather than a protease. Protease binding would transfer the wrong partner class, and the bounded term review did not establish another informative applicable binding term. The positive pair is retained.

Summary: Binding to IGFBP6 is retained as a non-core observation.

Reason: 3 exact source/partner record(s) were recovered from IntAct/MINT, with their assay methods and feature intervals preserved. HuRI is a binary assay in a heterologous system; it does not establish native colocalization or physiological interaction. IGFBP6 binds insulin-like growth factors; it is not itself an IGF or a growth-factor receptor. The positive binary pair is retained without inferring receptor, growth-factor or IGF-binding activity for EFEMP1.

Summary: Binding to OTX1 supports a verified partner-class refinement.

Reason: 3 exact source/partner record(s) were recovered from IntAct/MINT, with their assay methods and feature intervals preserved. HuRI is a binary assay in a heterologous system; it does not establish native colocalization or physiological interaction. OTX1 is a DNA-binding transcription factor. Binding to this partner supports the class term without transferring its developmental or transcriptional activities to EFEMP1.

Proposed replacements: DNA-binding transcription factor binding

Summary: Binding to HOXA1 supports a verified partner-class refinement.

Reason: 3 exact source/partner record(s) were recovered from IntAct/MINT, with their assay methods and feature intervals preserved. HuRI is a binary assay in a heterologous system; it does not establish native colocalization or physiological interaction. HOXA1 is a DNA-binding transcription factor. This refinement describes the class of the measured partner and does not assign DNA binding or transcriptional regulation to EFEMP1.

Proposed replacements: DNA-binding transcription factor binding

Summary: Binding to TRAF2 supports a verified partner-class refinement.

Reason: 3 exact source/partner record(s) were recovered from IntAct/MINT, with their assay methods and feature intervals preserved. HuRI is a binary assay in a heterologous system; it does not establish native colocalization or physiological interaction. TRAF2 has experimentally supported ubiquitin-protein ligase activity. The measured association supports ligase binding without implying that EFEMP1 catalyzes ubiquitination or is a ubiquitination substrate.

Proposed replacements: ubiquitin protein ligase binding

Summary: Binding to MKRN3 supports a verified partner-class refinement.

Reason: 3 exact source/partner record(s) were recovered from IntAct/MINT, with their assay methods and feature intervals preserved. HuRI is a binary assay in a heterologous system; it does not establish native colocalization or physiological interaction. MKRN3 has demonstrated ubiquitin-protein ligase activity. The binary interaction supports ligase binding, not EFEMP1 ubiquitination or endocrine regulation.

Proposed replacements: ubiquitin protein ligase binding

Summary: Binding to CLPP supports a verified partner-class refinement.

Reason: 3 exact source/partner record(s) were recovered from IntAct/MINT, with their assay methods and feature intervals preserved. HuRI is a binary assay in a heterologous system; it does not establish native colocalization or physiological interaction. CLPP is a peptidase with the conserved catalytic Ser153/His178 system in the matched parent sequence. Binding to a protease is supported; this does not establish substrate cleavage, mitochondrial import, or protease activity of EFEMP1.

Proposed replacements: protease binding

Summary: Binding to KRTAP19-2 is retained as a non-core observation.

Reason: 3 exact source/partner record(s) were recovered from IntAct/MINT, with their assay methods and feature intervals preserved. HuRI is a binary assay in a heterologous system; it does not establish native colocalization or physiological interaction. KRTAP19-2 is a hair keratin-associated matrix protein. Association with keratin does not itself place this protein within an intermediate filament, so cytoskeletal protein binding is not imposed on this positive pair.

Summary: Binding to KRTAP21-2 is retained as a non-core observation.

Reason: 3 exact source/partner record(s) were recovered from IntAct/MINT, with their assay methods and feature intervals preserved. HuRI is a binary assay in a heterologous system; it does not establish native colocalization or physiological interaction. KRTAP21-2 is a hair keratin-associated matrix protein. The inspected evidence does not establish that it is itself an actin-, microtubule- or intermediate-filament component; the positive interaction remains generic.

Summary: Binding to KRTAP6-1 is retained as a non-core observation.

Reason: 3 exact source/partner record(s) were recovered from IntAct/MINT, with their assay methods and feature intervals preserved. HuRI is a binary assay in a heterologous system; it does not establish native colocalization or physiological interaction. KRTAP6-1 is a hair keratin-associated matrix protein. Keratin association alone does not meet the definition requiring a cytoskeletal-component partner. The binary interaction is retained.

Summary: Binding to LCE1C is retained as a non-core observation.

Reason: 3 exact source/partner record(s) were recovered from IntAct/MINT, with their assay methods and feature intervals preserved. HuRI is a binary assay in a heterologous system; it does not establish native colocalization or physiological interaction. LCE1C is a cornified-envelope protein. This structural identity does not establish an actin-, microtubule- or intermediate-filament component, and no other informative supported binding class was identified.

Summary: Binding to LCE3A is retained as a non-core observation.

Reason: 3 exact source/partner record(s) were recovered from IntAct/MINT, with their assay methods and feature intervals preserved. HuRI is a binary assay in a heterologous system; it does not establish native colocalization or physiological interaction. LCE3A is a cornified-envelope protein rather than an established component of the cytoskeletal systems in GO:0008092. The positive binary observation is retained without transferring epidermal differentiation to EFEMP1.

Summary: Binding to LCE2C is retained as a non-core observation.

Reason: 3 exact source/partner record(s) were recovered from IntAct/MINT, with their assay methods and feature intervals preserved. HuRI is a binary assay in a heterologous system; it does not establish native colocalization or physiological interaction. LCE2C contributes to the cornified envelope; that does not by itself establish a cytoskeletal-component binding class. The interaction remains a positive non-core observation.

Summary: Binding to historical MEOX2 sequence Q6FHY5 supports a verified partner-class refinement.

Reason: 3 exact source/partner record(s) were recovered from IntAct/MINT, with their assay methods and feature intervals preserved. HuRI is a binary assay in a heterologous system; it does not establish native colocalization or physiological interaction. The 303-residue Q6FHY5 sequence matches the recorded checksum and is exactly current MEOX2/P50222 with canonical Q81 deleted; the homeobox remains. This supports transcription-factor binding while retaining the literal imported partner identifier. It does not establish native interaction or EFEMP1 transcriptional activity.

Proposed replacements: DNA-binding transcription factor binding

Summary: Binding to MSRB3 product 2 supports a verified partner-class refinement.

Reason: 3 exact source/partner record(s) were recovered from IntAct/MINT, with their assay methods and feature intervals preserved. HuRI is a binary assay in a heterologous system; it does not establish native colocalization or physiological interaction. The exact 185-residue product-2 sequence matches the record checksum. Its alternative N terminus retains the MsrB domain and catalytic cysteine (canonical C158, product-2 C151), supporting enzyme-class binding. Neither EFEMP1 reductase activity nor a native mitochondrial interaction is inferred.

Proposed replacements: enzyme binding

Summary: Binding to NR4A3 supports a verified partner-class refinement.

Reason: 3 exact source/partner record(s) were recovered from IntAct/MINT, with their assay methods and feature intervals preserved. HuRI is a binary assay in a heterologous system; it does not establish native colocalization or physiological interaction. NR4A3 is a DNA-binding nuclear receptor transcription factor. The informative claim is binding to that factor, not receptor signaling or DNA binding by EFEMP1.

Proposed replacements: DNA-binding transcription factor binding

Summary: Binding to FAM110A is retained as a non-core observation.

Reason: 3 exact source/partner record(s) were recovered from IntAct/MINT, with their assay methods and feature intervals preserved. HuRI is a binary assay in a heterologous system; it does not establish native colocalization or physiological interaction. The inspected FAM110A record supports centrosomal localization but does not establish a cytoskeletal structural-component or enzyme class. Localization alone is insufficient to refine this positive binding observation.

Summary: Binding to LCE3D is retained as a non-core observation.

Reason: 3 exact source/partner record(s) were recovered from IntAct/MINT, with their assay methods and feature intervals preserved. HuRI is a binary assay in a heterologous system; it does not establish native colocalization or physiological interaction. LCE3D is a cornified-envelope protein; this does not itself satisfy the actin/microtubule/intermediate-filament component definition. The positive pair is retained without an invented subclass.

Summary: Binding to HSD3B7 supports a verified partner-class refinement.

Reason: 3 exact source/partner record(s) were recovered from IntAct/MINT, with their assay methods and feature intervals preserved. HuRI is a binary assay in a heterologous system; it does not establish native colocalization or physiological interaction. HSD3B7 is an experimentally characterized sterol dehydrogenase; its matched sequence retains the annotated catalytic Tyr159. The interaction supports enzyme binding, not sterol catalysis by EFEMP1.

Proposed replacements: enzyme binding

Summary: Binding to ANAPC11 product 2 is retained as a non-core observation.

Reason: 3 exact source/partner record(s) were recovered from IntAct/MINT, with their assay methods and feature intervals preserved. HuRI is a binary assay in a heterologous system; it does not establish native colocalization or physiological interaction. The matched 196-residue product replaces canonical residues 38–84, encompassing most of the RING domain (34–77). The E3-ligase class of canonical ANAPC11 cannot be assumed for this structurally altered product. Positive binding is preserved without asserting catalytic inactivity or a narrower class.

Summary: Binding to UBQLN2 is retained as a non-core observation.

Reason: 3 exact source/partner record(s) were recovered from IntAct/MINT, with their assay methods and feature intervals preserved. HuRI is a binary assay in a heterologous system; it does not establish native colocalization or physiological interaction. UBQLN2 is a protein-quality-control shuttle/adaptor. Its involvement in degradation does not make it a protease or ubiquitin ligase, and the inspected evidence does not establish a more informative applicable class for this pair.

Summary: Binding to BAG5 is retained as a non-core observation.

Reason: 3 exact source/partner record(s) were recovered from IntAct/MINT, with their assay methods and feature intervals preserved. HuRI is a binary assay in a heterologous system; it does not establish native colocalization or physiological interaction. BAG5 regulates HSP70 as a co-chaperone/nucleotide-exchange factor. That does not alone establish the direct nascent/unfolded-protein-binding chaperone class required for GO:0051087; the positive association is retained.

Summary: Binding to HTT is retained as a non-core observation.

Reason: 3 exact source/partner record(s) were recovered from IntAct/MINT, with their assay methods and feature intervals preserved. The interaction records specify HTT sufficient binding region 1–546 and polyglutamine-region mutations. No informative supported partner-class term follows from this fragment, and the result does not establish an EFEMP1 neurodegeneration or transport mechanism.

Summary: Secreted EFEMP1 is consistent with the human colostrum annotation.

Reason: The source abstract describes depletion of abundant proteins and proteomic identification of minor proteins in the aqueous phase of human colostrum. Its EFEMP1-specific table was not recovered. Extracellular localization is independently demonstrated for human EFEMP1, so the curator-supported broad location is retained without claiming independent inspection of the colostrum target measurement.

Summary: Human aortic extracellular fractions contain fibulin-3.

Reason: The original fractionation and mass-spectrometry tables explicitly identify FBLN3_HUMAN through multiple unique peptides. This directly supports an extracellular pool without assigning a specific natural splice product or proving exclusive extracellular localization.

Summary: Human venous extracellular proteomics is consistent with the established secreted pool.

Reason: The original compares normal and varicose human saphenous veins using salt/detergent/guanidine extraction and mass spectrometry. The recovered supplementary package contains figures rather than the EFEMP1-specific protein table. The broad extracellular assignment is retained with curator deference and independent human secretion/matrix evidence, not presented as a target row independently inspected in this study.

Summary: Human EFEMP1 is secreted in the transfected-cell assay.

Reason: The original uses a human EFEMP1 ORF with a C-terminal triple-FLAG tag in mouse ATDC5 cells and detects tagged protein in conditioned medium. Purified human tagged protein was also added extracellularly. Host species is not used to rename the tested human protein, and the construct is not assigned to a natural splice product.

Summary: The curated Reactome entity places mature human EFEMP1 extracellularly.

Reason: The saved current chain is 1592387 input DefinedSet 2395327 hasMember 2161393. Entity 2161393 explicitly maps to human UniProt Q12805 residues 18–493 in the extracellular region. This is exact model membership, not an inference from another fibulin or a claim of an independently measured binding interface.

Summary: The curated Reactome entity places mature human EFEMP1 extracellularly.

Reason: The saved current chain is 2161282 input complex 2396077 → complex 2395337 → DefinedSet 2395327 hasMember 2161393. Entity 2161393 explicitly maps to human UniProt Q12805 residues 18–493 in the extracellular region. This is exact model membership, not an inference from another fibulin or a claim of an independently measured binding interface.

Summary: HPA reports a positive mitochondrial staining pool.

Reason: The target HPA page individually rates Mitochondria as enhanced, while the XML overall reliability is supported. Antibody/cell-line records contain mitochondrial staining, with additional nuclear pools in some contexts. This positive intracellular observation is retained as non-core without treating the secretion signal as disproof. The images were not independently scored, and an isoform or mitochondrial biochemical function is unresolved.

Summary: Steady-state marker changes accompany altered chondrogenic differentiation; transcription regulation itself was not measured.

Reason: Stable expression of tagged human EFEMP1 in mouse ATDC5 cells alters differentiation and steady-state transcript abundance measured by reverse-transcription quantitative PCR, normalized to beta-actin (Methods, journal page 1117; Figures 3–4, page 1120). These endpoints do not measure transcription rate, nascent RNA, promoter activity or another defined transcription-regulatory step. The separate extracellular-addition experiment measures cartilage aggregates and proteoglycan staining (Figure 2C–D); it is not the source of the Sox marker-transcript results. The transcription-specific assertion is therefore over-annotated relative to the measured differentiation and RNA-abundance endpoints. This judgment follows the inspected original Methods and Results, not abstract-only access or a requirement for direct DNA binding; the source assertion and contextual differentiation effect remain preserved.

Summary: EFEMP1 stimulates EGFR-dependent signaling in carcinoma cells.

Reason: The recovered primary study measures EGFR/Akt/ERK phosphorylation after purified human EFEMP1 treatment, with pathway-inhibitor context. The positive signaling effect is retained as a context-dependent function in the tested pancreatic carcinoma system; it is not generalized to every normal tissue or splice product.

Summary: The retinal disease association does not resolve the exact visual-perception process claim.

Reason: The verified abstract identifies EFEMP1 Arg345Trp in families with inherited drusen-associated retinal disease. GO:0007601 concerns reception, conversion and recognition of visual stimuli. The original body supporting this TAS assertion was not recovered, so a specific contribution to that process cannot be adjudicated from the disease association alone. Retinal disease relevance is retained without either asserting phototransduction or rejecting an uninspected experiment.

Summary: The measured tyrosine phosphorylation is performed by EGFR.

Reason: Purified EFEMP1 treatment induces EGFR phosphorylation in the human carcinoma assay. EFEMP1 is the extracellular interacting factor, whereas the receptor kinase performs tyrosine phosphorylation. The observed effect is real but does not establish EFEMP1 itself executing peptidyl-tyrosine phosphorylation; binding and contextual signaling retain the supported biology.

Summary: Human aortic proteomics directly detects fibulin-3 in extracellular matrix preparations.

Reason: Original Tables I–III identify FBLN3_HUMAN with multiple unique peptides after sequential tissue extraction. The ordinary source relation is preserved without driving a separate biological verdict, and the protein-level detection does not resolve natural splice products.

Summary: EFEMP1 is detected in human colon/liver matrix-enriched preparations.

Reason: The original sequential extraction and recovered Supplementary Table S2 identify EFEMP1 in normal and tumor-associated matrix contexts. The full multi-accession group is preserved and does not uniquely assign a splice product. Distinct normal/tumor tissue extensions are not collapsed into an invented tissue-specific mechanism.

Summary: The curated human matrix-proteomics location is consistent with independent target evidence.

Reason: The original describes sequential cellular-component removal followed by ECM solubilization and mass spectrometry in human breast/omental samples. The target-specific supplementary table was not recovered; the broad matrix location is retained with curator deference and independently recovered EFEMP1 matrix measurements, without claiming a source-specific peptide count.

Summary: The ancestral matrix-location inference is supported in the EFEMP1 lineage.

Reason: Human EFEMP1 lies below the literal PTN000620536 IBD in the current PTHR24039 tree, and its aligned sequence matches the official reference. Current IBD and imported GOA donor lists differ, so both snapshots are retained rather than rewriting the source. Curated experimental descendant locations and direct human EFEMP1 matrix evidence support the broad inherited compartment; no fibrillin-specific architecture is transferred.

Summary: Human EFEMP1 supports extracellular matrix network organization.

Reason: Human ARPE-19 EFEMP1 knockout causes loss of normal collagen VI network organization in decellularized matrices; wild-type expression partly rescues, whereas the C55R variant does not. Extracellular matrix and conditioned-medium measurements support the location. MMP2 activity changes do not make EFEMP1 a protease, and mutant disulfide dimers do not establish obligate wild-type dimerization.

Summary: The broad extracellular matrix assignment is independently established.

Reason: The source abstract places the EFEMP1 mutation in inherited retinal disease; its full body was unavailable. Broad matrix localization is nevertheless directly supported by later human fractionation, matrix proteomics and retinal-cell matrix experiments. The TAS assertion is retained without presenting the clinical abstract as an independent matrix-localization assay.

Summary: Human EFEMP1 suppresses chondrocyte differentiation in a mouse-cell assay.

Reason: Stable expression of tagged human EFEMP1 in mouse ATDC5 cells suppresses chondrogenic differentiation and changes cartilage-marker and Sox transcript abundance (Figures 3–4). In a separate experiment, extracellular addition of purified tagged human EFEMP1 produced in Sf9 cells reduces cartilage aggregates and proteoglycan staining (Figure 2C–D); the marker-transcript measurements were not assigned to that addition arm. Both experiments support a contextual negative effect on differentiation, without establishing a direct human cartilage-development experiment. The proposed TIMP3/ADAM mechanism remains an untested Discussion hypothesis.

Summary: The cited expression study assays fibrillins, not EFEMP1/fibulin-3.

Reason: The full original Methods specify antibodies to FBN1, FBN2 and FBN3, and Results compare those fibrillin isoforms and MAGP1 in human embryonic/fetal and juvenile eyes. The target called fibrillin-3 is FBN3, not EFEMP1 (fibulin-3). This verified target mismatch in the operative paper does not support the imported EFEMP1 IEP assertion. The original reference identifier and source dictionary are retained; the conclusion is not based on the title or abstract alone.

Summary: The cited expression study assays fibrillins, not EFEMP1/fibulin-3.

Reason: The full original Methods specify antibodies to FBN1, FBN2 and FBN3, and Results compare those fibrillin isoforms and MAGP1 in human embryonic/fetal and juvenile eyes. The target called fibrillin-3 is FBN3, not EFEMP1 (fibulin-3). This verified target mismatch in the operative paper does not support the imported EFEMP1 IEP assertion. The original reference identifier and source dictionary are retained; the conclusion is not based on the title or abstract alone.

Summary: The cited expression study assays fibrillins, not EFEMP1/fibulin-3.

Reason: The full original Methods specify antibodies to FBN1, FBN2 and FBN3, and Results compare those fibrillin isoforms and MAGP1 in human embryonic/fetal and juvenile eyes. The target called fibrillin-3 is FBN3, not EFEMP1 (fibulin-3). This verified target mismatch in the operative paper does not support the imported EFEMP1 IEP assertion. The original reference identifier and source dictionary are retained; the conclusion is not based on the title or abstract alone.

Summary: The EFEMP1-specific exosome measurement remains unverified.

Reason: The original describes differential-ultracentrifugation EPS-urine vesicle preparations, a sucrose/DTT wash and mass spectrometry with less than 1% FDR. The EFEMP1 target row in Supplementary Table 2 was not recovered: attempted downloads returned non-workbook content. The general preparation and EFEMP1 secretion do not settle specific vesicle association. Neither contamination nor exclusion from exosomes is asserted.

Summary: Elastic-fiber association is consistent with independently accessible source evidence.

Reason: The cited current ARBA rule was retrieved, but its complete triggering condition could not be reconstructed from the current target cross-reference snapshot. This lineage/access limitation does not refute the biology: the recovered elastic-fiber review explicitly includes fibulin-3 among fiber-associated molecules, and the curated Reactome model contains EFEMP1 in the fibulin member set. The IEA location is retained without transferring all fibulin-4/5 binding mechanisms.

Summary: The cited review explicitly identifies fibulin-3 as elastic-fiber associated.

Reason: The recovered publisher text names small fibulins 3–5 among elastic-fiber-associated molecules and describes tissue-specific elastic-fiber defects in fibulin-3-null mice. This supports the location at the stated conserved scope. Detailed tropoelastin/LOX binding results for fibulin-4/5 are not reassigned to EFEMP1, and equivalent localization of all human splice products is not assumed.

Summary: The specific basement-membrane assignment remains unresolved.

Reason: The current ARBA rule was inspected, but its triggering target condition was not fully reproduced. The term specifies the non-collagenous solid-phase compartment of basement membrane, which is more specific than secretion or general matrix detection. Available disease-associated subretinal accumulation and broad matrix evidence do not by themselves resolve that exact compartment. The inference is retained for adjudication without asserting absence from basement membranes.

Summary: Non-collagenous interstitial-matrix membership is consistent with the supported matrix pool.

Reason: MatrisomeDB 2.0 integrates uniformly reprocessed human/mouse matrix-enriched proteomics; its recovered supplement did not supply an EFEMP1-specific compartment row. Independent human matrix measurements and the cited elastic-fiber review support a non-collagenous glycoprotein pool. Current typed ontology relationships place elastic fiber within the non-collagenous component of interstitial matrix. This is complementary corroboration, not a claim that the database paper directly imaged EFEMP1 in every interstitial tissue.