EFEMP1 is a secreted extracellular matrix glycoprotein. Human retinal-cell loss/rescue experiments establish a contribution to extracellular network integrity; a separate human pancreatic carcinoma study establishes EGFR association and receptor-dependent signaling. These mechanisms are represented as two qualified core activities. Neither supports EFEMP1 kinase/protease activity, and the matrix phenotype is not a purified collagen-binding assay. PMID:35998264(https://pubmed.ncbi.nlm.nih.gov/35998264/), PMID:19804359(https://pubmed.ncbi.nlm.nih.gov/19804359/).
In human ARPE-19 cells, EFEMP1 loss disrupts collagen VI network organization in decellularized matrix and reduces extracellular MMP2 activity. Wild-type EFEMP1 partly rescues the matrix phenotype, whereas the C55R variant fails despite secretion. The measured disulfide-linked dimers concern the mutant; they do not establish obligatory wild-type covalent dimers. Changes in MMP2 activity are an extracellular outcome, not evidence that EFEMP1 itself is a protease. PMID:35998264(https://pubmed.ncbi.nlm.nih.gov/35998264/).
Human aortic fractionation and human colon/liver matrisome experiments independently identify fibulin-3 with multiple peptides. The colon/liver workbook retains a mixed accession group, including several splice-product identifiers; shared peptide evidence cannot assign a natural isoform. RCA classification as a matrix structural constituent is supported by the term's scope and the separate human knockout/rescue evidence, rather than being presented as a mechanical-strength measurement. The reproducible extract preserves the target measurements. PMID:20551380(https://pubmed.ncbi.nlm.nih.gov/20551380/), PMID:25037231(https://pubmed.ncbi.nlm.nih.gov/25037231/).
Human carcinoma-cell co-immunoprecipitation and EGF competition support EFEMP1–EGFR association. Purchased human EFEMP1 stimulates EGFR and downstream Akt/ERK phosphorylation in the measured system. The paper does not establish a purified binary interface, an identified natural splice product, or the same signaling mechanism in all normal tissues. EFEMP1 is the extracellular interacting factor; EGFR performs the phosphorylation. Accordingly, receptor activity is refined to receptor binding, and the direct phosphorylation-process assertion is over-annotated. The positive carcinoma signaling observation remains non-core. PMID:19804359(https://pubmed.ncbi.nlm.nih.gov/19804359/).
Human EFEMP1 expression and extracellular addition of purified tagged human protein inhibit chondrogenic differentiation in mouse ATDC5 cells. Conditioned-medium detection supports secretion. Reduced cartilage-like nodules, proteoglycan accumulation and marker expression support a contextual differentiation effect, not direct human cartilage-development measurements or an EFEMP1 DNA-binding mechanism. The human protein, mouse assay host and insect-cell production host are distinct. The suggested TIMP3/ADAM mechanism was not tested in this experiment; the separate TIMP3 interaction abstract does not close that mechanistic gap. PMID:20005202(https://pubmed.ncbi.nlm.nih.gov/20005202/), PMID:15123717(https://pubmed.ncbi.nlm.nih.gov/15123717/).
Positive screen interactions are retained with their assay limits. Informative refinements describe the verified class of the measured partner, without transferring its enzymatic or transcriptional activity to EFEMP1. A binding site need not be the partner's catalytic domain: the recorded ARAF binding regions are N-terminal regulatory segments, while ARAF remains a kinase protein. Conversely, a canonical gene's activity is not automatically assigned to an altered product whose relevant domain is replaced. The partner assessments and source guide expose the important sequence and feature distinctions.
The current interaction-query extract preserves 46 literal source/partner joins and all fields of matching records. Reference-sequence checksums identify the parent sequence associated with a record; they do not prove the experimental insert was full length. Separately annotated binding regions, variants and processed chains remain explicit. HuRI observations are binary interactions in a heterologous system, not proof of native colocalization.
Supported generic interactions without an evidence-backed narrower class remain KEEP_AS_NON_CORE. This follows the intentional generic-binding exception in the ClinGen project instructions, rather than rejecting positive interactions as false or inventing a partner class. Ordinary source qualifiers are preserved and do not independently change these verdicts.
Current PANTHER responses place EFEMP1 in PTHR24039, while the older machine record points to PTHR24050. The resolved 330-node tree and alignment contain the exact 493-residue target below the literal ancestral node PTN000620536. Both IBD assertions, experimental descendant annotations and the target path support the broad matrix activities. Current IBD and imported GOA donor lists differ; both snapshots are preserved. Donor number and evidence on a descendant are not circularity defects.
HPA reports mitochondrial staining with location-specific enhanced reliability, whereas its XML overall score is supported. This positive intracellular observation is retained as non-core. The secretion signal does not disprove an intracellular pool; no mitochondrial biochemical function or splice-product identity is inferred, and the images were not independently scored. The exact HPA location and antibody/cell-line context are linked in the source guide.
The elastic-fiber review explicitly includes fibulin-3 among associated proteins and discusses tissue-specific defects in fibulin-3-null mice. Detailed fibulin-4/5 tropoelastin/LOX mechanisms are not transferred. Reactome independently contains mature human EFEMP1 as a member of its fibulin set. Current typed ontology relationships place elastic fiber within the non-collagenous component of interstitial matrix. The narrower basement-membrane inference remains unresolved because the complete ARBA trigger and exact compartment were not independently established. PMID:23962539(https://pubmed.ncbi.nlm.nih.gov/23962539/).
The EFEMP1 Arg345Trp association with inherited drusen-associated retinal disease is supported by human family genetics. That disease relationship alone does not resolve the precise visual-perception process, which concerns reception and processing of visual stimuli; the original body supporting that assertion was unavailable. Broad matrix localization is independently established. PMID:10369267(https://pubmed.ncbi.nlm.nih.gov/10369267/).
Three eye-development expression assertions cite a study of FBN1, FBN2 and FBN3. Its full Methods specify fibrillin antibodies, and its Results concern fibrillins and MAGP1 in human eyes. Fibrillin-3 is FBN3, not EFEMP1/fibulin-3. These three assertions are removed on verified operative-source identity, while the original source objects and PMID remain intact. This is not an inference from an inaccessible title or abstract. PMID:25406291(https://pubmed.ncbi.nlm.nih.gov/25406291/).
The exosome assertion remains UNDECIDED: the original vesicle preparation and proteomics methods were read, but the EFEMP1 target row in Supplementary Table 2 was not recovered. A nominal workbook download returned a challenge document. Neither contamination nor exclusion from extracellular vesicles is asserted. PMID:23533145(https://pubmed.ncbi.nlm.nih.gov/23533145/).
The five official products reconstruct to 493, 485, 492, 492 and 355 residues. Product 5 lacks the canonical signal region and two central EGF modules. Equivalent secretion, stability, matrix function and receptor signaling have not been demonstrated. The older cloning paper reports 487 residues; without its construct sequence this is not equated with one of the current products. Its abstract's electron-microscopy result concerns fibulin-4. PMID:10601734(https://pubmed.ncbi.nlm.nih.gov/10601734/).
All 79 imported source dictionaries are preserved: 21 ACCEPT, 29 MODIFY, 22 KEEP_AS_NON_CORE, three UNDECIDED, one MARK_AS_OVER_ANNOTATED and three REMOVE. No new annotation or quotation is added. The 81 intake QuickGO contexts represent 79 unique imported tuples; two matrix-proteomics tuples have distinct intestine/liver extension contexts, retained in the source provenance. No source retirement is needed. Automated research attempts failed authentication/quota checks; the evidence assessment instead records its actual primary-source and database access. Downloaded publication text and machine source flags remain unchanged.
The source-specific binding review preserves all 27 partner-class refinements. Literal records and independently replayed reference sequences support the assayed pair identities; 21 of the 27 include validated-two-hybrid records. The six others retain their actual methods and access limits. Retesting, dataset-level orthogonal validation and native physiological evidence are distinguished in the source guide. No natural extracellular interaction, all-product equivalence, catalytic-domain contact or EFEMP1 enzyme/transcription activity is inferred from a partner class. The different chaperone-binding decisions follow the inspected partner classes, not unequal treatment of the same assay.
This paragraph supersedes the earlier combined description of the two PMID:20005202 experimental arms and the earlier action tally. Stable tagged-human-EFEMP1 expression in mouse ATDC5 cells produced the steady-state marker-transcript results (Figures 3–4); the separate extracellular-addition experiment measured aggregate formation and proteoglycan staining (Figure 2C–D). The complete operative Methods and Results were independently examined, including visual checks of Methods and Figures 3–4. These endpoints support contextual inhibition of differentiation but do not directly assay transcription regulation. Accordingly, the transcription row is now MARK_AS_OVER_ANNOTATED, while the differentiation row remains KEEP_AS_NON_CORE. The normal abstract-only cache and the separately recovered full original are distinguished in the source guide and reference assessment.
All 79 imported source dictionaries, partner identities, five products, two cores and existing excerpts remain unchanged. The resulting dispositions are 21 ACCEPT, 29 MODIFY, 21 KEEP_AS_NON_CORE, three UNDECIDED, two MARK_AS_OVER_ANNOTATED and three REMOVE. No annotation row or quotation is added.
The nine reviewed ATN1, HOXA1, MEOX2, OTX1, CLPP and NR4A3 binding assertions retain their supported partner-class refinements. Their source-linked evidence consists of yeast two-hybrid assays, including recorded fragments and retests. These assays establish the reported interaction in that experimental setting; they do not demonstrate an endogenous human complex or physiological colocalization. The secretion of EFEMP1 and the predominant nuclear or mitochondrial locations of several partners therefore remain relevant uncertainties, without establishing that the positive assays are false. Parent-sequence checksums identify reference proteins, not every experimental insert or mature processed product.
The ClinGen project binding policy uses an informative replacement when the measured partner and its class are supported, and retains supported generic binding as non-core when a specific replacement is unestablished. Applying it here does not give EFEMP1 transcription-factor, corepressor or protease activity. The source guide separates the assay, construct, class and native-context limits for each of these nine assertions. Four older-source reasons now expose these limits directly; the five HuRI reasons already contained them. No annotation action or core function changes.
The three PMID:25406291 removals retain their verified fibrillin-versus-fibulin target distinction. The published abstract cache and full original are distinguished in the guide, with exact cache hashes, line locators, and the operative antibody-panel and specificity-control sections. The committed cache remains abstract-only; the full Methods were read separately. No new quotations or cache replacements are needed to expose this evidence.