This guide links the underlying evidence and distinguishes measured observations from interpretation. The review YAML contains the decisions; this guide is not an independent primary source. The evidence extract reproduces selected records from hash-identified raw inputs without bundling full articles. Reference access flags describe operative access, not simply the normal publication cache's full_text_available metadata.
Human matrix contribution — PMID:35998264. The original article was available in the independently fetched full publication cache, including the human cell experiments. A separate XML retrieval failed and is not treated as evidence. Methods specify human ARPE-19 loss/rescue and tagged human constructs in human production cells; the tag is inserted after the signal peptide. Results show altered collagen VI matrix organization, extracellular MMP2 activity and C55R mutant behavior. Wild-type partial rescue and mutant failure support matrix contribution. No purified collagen-binding assay, EFEMP1 protease chemistry or equivalence among natural products is inferred. Selected Methods and all operative Results were read; supplements were not claimed as inspected.
EGFR association — PMID:19804359. The institutional record supplies the publisher PDF, SHA256 5adb48b843728739ec551b36328bd6c2bed6c690101307528a6161c2301bb800. Selected Methods/Results and PDF page 4 figures 4–7 establish human FG/L3.6pl lysate co-immunoprecipitation, the no-antibody control, reduced association after EGF treatment, and signaling after purchased human EFEMP1. The supplier product is not identified by an accession or construct sequence. An 823-base cDNA check addresses the Arg345 region, not the full assayed protein. The paper contains inconsistent receptor phosphosite labels; no harmonized residue claim is made. Normal cached text remains abstract-only despite this separate original access.
Chondrogenesis and secretion — PMID:20005202. A public research-report PDF contains the article on physical pages 2–7, journal pages 1116–1121. PDF SHA256 is e8354cc8fbbf70c4c198eed1e6e807d62989812647445782de1e0f1c8bf37529. The original was read through page-specific OCR, including operative Methods, Results and Discussion; visual checks include Methods and Figures 3–4. Human IMAGE clone 6083389 was expressed with a C-terminal triple-FLAG tag in mouse ATDC5 cells; purified tagged human protein was produced using Sf9 cells. Stable-expression clones show differentiation and steady-state transcript changes measured by reverse-transcription quantitative PCR normalized to beta-actin (Methods, journal page 1117; Figures 3–4, page 1120). Separately, extracellular addition reduces cartilage aggregates and proteoglycan staining (Figure 2C–D). The Sox marker-transcript measurements belong to the stable-expression experiment. Neither experiment measures transcription rate, nascent RNA or promoter activity. The construct is not assigned to a current natural product. The proposed TIMP3/ADAM mechanism remains an untested interpretation.
Cloning and TIMP3 sources — PMID:10601734 and PMID:15123717. Verified abstracts support fibulin-3 cloning/extracellular distribution and the TIMP3 interaction, respectively. Full original bodies were not recovered. The historical 487-residue cloning sequence is not reconciled to the current 493-residue reference. The microscopy statement in the cloning abstract concerns fibulin-4. The TIMP3 abstract's disease-associated deposits are not a complete normal-compartment experiment or proof of an enzymatic mechanism.
| Source | Actual access and target evidence | Limit retained |
|---|---|---|
| PMID:20551380(https://pubmed.ncbi.nlm.nih.gov/20551380/) | Original human aorta Methods and Tables I–III; FBLN3_HUMAN target fields reproduced in records.json |
Sequential extraction and peptide detection do not resolve natural products or mechanical strength |
| PMID:25037231(https://pubmed.ncbi.nlm.nih.gov/25037231/) | Original Methods and Additional File 2; EFEMP1 physical row 43 in complete-data and matrisome sheets | Mixed accession group and multiple tissue contexts are preserved; no isoform-specific assignment |
| PMID:27068509(https://pubmed.ncbi.nlm.nih.gov/27068509/) | Human normal/varicose saphenous-vein Methods/Results; recovered supplement contains figures | EFEMP1-specific protein table unavailable; broad extracellular location has separate human evidence |
| PMID:28675934(https://pubmed.ncbi.nlm.nih.gov/28675934/) | Human breast/omental ECM-enrichment Methods, filtering and Discussion in the independently fetched full text | Target supplementary row unavailable; no invented peptide count or direct mechanical assay |
| PMID:16502470(https://pubmed.ncbi.nlm.nih.gov/16502470/) | Human colostrum abstract | Full body/target table unavailable; broad extracellular pool retained with curator deference and independent evidence |
| PMID:23533145(https://pubmed.ncbi.nlm.nih.gov/23533145/) | Main original vesicle-preparation and proteomics methods | Target Supplementary Table 2 unavailable; the nominal XLSX route served a challenge, so exosome association remains unresolved |
| PMID:36399478(https://pubmed.ncbi.nlm.nih.gov/36399478/) | MatrisomeDB original Methods/Results and supplementary workbooks | No target compartment row in the recovered supplement; location is corroborated by independent matrix/elastic-fiber evidence |
HPA EFEMP1 subcellular data report mitochondria with location-specific enhanced reliability; the XML overall reliability is supported. The antibody/cell-line records include additional nuclear pools in some contexts. These are positive observations, not a demonstrated mitochondrial biochemical role. Images were not independently scored. The extracellular signal sequence is not used as a topological disproof of this pool.
The elastic-fiber review (PMID:23962539) explicitly discusses fibulin-3 association and tissue-specific mouse defects. Its fibulin-4/5 binding mechanisms remain separate. Current elastic-fiber GO relationships support the broader non-collagenous interstitial matrix context. They do not establish the distinct basement-membrane compartment.
records.json retains the complete fields of every matching current IntAct/MINT record for the 46 imported source/partner joins. Its source_index is zero-based and paired with literal PMID/partner identifiers, so row numbering is never the sole evidence key. The two current-query absences are not treated as negative experiments:
For PMID:12620389, 16189514, 16713569 and 32814053, abstracts plus literal source-specific interaction records were available, but original bodies were not recovered. PMID:25416956, 25910212 and 26871637 have incomplete accessible main-text sections; their database records provide exact pair/variant/feature evidence, without a claim to unseen complete Methods or inserts. Selected operative text was also read for PMID:21653829, 31515488 and 32296183. HuRI's heterologous binary-assay context remains explicit.
The source-linked partner-class table accompanies the literal records. Important distinctions are:
The 27 source-specific partner-class refinements retain the positive assay assertions with their limitations. Independent replay of the literal PMID/partner joins finds 21 with an IntAct/MINT MI:1356 validated-two-hybrid record. The other six are ARAF/PMID:12620389, SGTA/PMID:16189514, ATN1/PMID:16713569, HOXA1/PMID:21653829 and the SGTA and BAG6 pairs from PMID:31515488. They retain their recorded two-hybrid, pooling or array methods rather than being relabeled validated-two-hybrid. Every exact stable interaction ID, species field, feature interval and method remains available in the literal records. Multiple screening stages or repeated database entries are not independent biological replications.
| Study | Measured scope and limits |
|---|---|
| 12620389 | Human RAF N-terminal regulatory-domain Y2H; ARAF records EBI-537167, EBI-536929 and EBI-537034 retain their three intervals. The abstract and curated pair records were inspected; no full-body or kinase-active-site contact claim. |
| 16189514 | SGTA pooling record EBI-759046 / IM-16520-2291. The abstract describes stringent binary screening and dataset-level co-affinity validation, not a verified orthogonal result for this particular pair. |
| 16713569 | Three ATN1 Y2H records retain their distinct fragment intervals. The abstract's validation percentage concerns a selected set; pair-specific mammalian validation is not asserted. |
| 21653829 | HOXA1 record EBI-26512440 / IM-28561-341. Results describe independent reconstitution testing of retained Y2H pairs and a selected mammalian-validation subset; membership of this pair in that subset is unverified. Its 1–110 feature is separate from the parent's homeobox and sequence checksum. |
| 25416956 | Five exact pair joins have validated, array and pooling records. Accessible main text is incomplete; method codes and insert features come from the literal curated records, not an invented complete Methods read. |
| 25910212 | SGTA, MEOX2 and TRAF2 joins include validated, array and pooling records, with EFEMP1 residue-345 mutation-with-no-effect features. These do not establish natural-product equivalence or a native interaction mechanism. Accessible main text is incomplete. |
| 26871637 | Three validated/array/pooling joins retain reference-checksum and feature scope. The study's isoform-comparison design does not establish that all five current EFEMP1 products share the interaction. Accessible main text is incomplete. |
| 31515488 | SGTA and BAG6 array records preserve the side-specific variant features. Methods describe paired WT/mutant Y2H with empty-vector autoactivation controls and reproducibility criteria. Orthogonal PCA was performed on a subset; these two pairs are not claimed individually PCA-confirmed. |
| 32296183 | All ten disputed HuRI refinements have validated/array/pooling records. Results describe quadruplicate pairwise retesting and sequence confirmation. Orthogonal-assay benchmarking is a dataset-level result, not proof for every EFEMP1 pair. Human processing, native colocalization and all isoforms remain unresolved. |
The binding terms describe interaction with a class of partner; they do not assign that partner's catalysis or transcriptional function to EFEMP1. Their definitions do not require recognition of a catalytic site, protease inhibition, or a demonstrated folding-client mechanism. The secretion signal and disulfide-rich architecture make native context an important unresolved question, but do not by themselves demonstrate that these specific positive assays arose from misfolding. The revised judgments neither infer an extracellular interaction with every partner nor claim the results are false.
For protein-folding chaperone binding, SGTA and BAG6 have curated direct client-binding/antiaggregation roles. RIC8A's nascent-G-alpha folding classification is explicitly supported by ortholog evidence in its reviewed record; the human activation experiment is not presented as a human folding assay. Its product-4 reference sequence differs by deletion of residue 209, while the recorded 111–530 fragments remain distinct from full-length insert proof. BAG4/BAG5 regulate HSP70 and UBQLN2 connects ubiquitinated substrates with degradation machinery; those inspected assignments alone do not establish the direct nascent/unfolded-polypeptide-binding class required here. Their positive generic interactions therefore remain non-core. This is a class-evidence distinction, not a claim that the same screen is reliable for one partner and invalid for another.
The portable extract records the exact target path and two literal IBD assertions from PTN000620536. The current family is PTHR24039, while the older PTHR24050 pointer remains unchanged in the machine source. The target MSA sequence matches Q12805 exactly; this authenticates the reference sequence, not an experimental insert. Current matrix-location IBD donors differ from the imported GOA snapshot and are retained separately. Experimental descendant records include human fibrillin and worm fibulin evidence; donor paper bodies were not claimed as independently read. No relevant NOT/IRD was found on the target path. Node placement, rather than donor count, grounds the transfer.
The current InterPro IPR001881 and IPR018097 mappings support calcium-binding EGF modules. This is domain inference, not a measured EFEMP1 calcium affinity. Current ARBA00084594 and ARBA00096893 rule bodies were inspected, but complete triggering conditions were not reconstructed from the target snapshot. Independently supported elastic-fiber location is retained; the narrower basement-membrane assignment remains unresolved.
The exact Reactome chains are:
hasMember 2161393.hasMember 2161393.Entity 2161393 identifies mature human Q12805 residues 18–493 in the extracellular region. Set membership supports the modeled location; it is not a purified physical-interaction assay.
For PMID:25406291, complete operative Methods/Results and Discussion describe FBN1/FBN2/FBN3 and MAGP1 in developing human eyes. Methods specify fibrillin antibodies, not EFEMP1. The three IEP removals are based on that verified target distinction, not a title-only inference. The publication identifier and original annotation fields remain unchanged.
The three unresolved assertions concern visual perception, the target-specific exosome measurement, and the specific basement-membrane compartment. Human retinal disease, secretion and general matrix detection do not by themselves settle these narrower claims. Positive observations and source provenance remain intact.
The following nine source-specific refinements are supported by the literal interaction records and partner-class assessments. The 25 matched records include screening stages, retests and fragments; they are not 25 independent biological replications. Both participants are recorded as human proteins assayed in yeast. The terms transcription coregulator binding, DNA-binding transcription factor binding and protease binding describe the measured partner class, without assigning its activity to EFEMP1.
| Source and partner | Measured pair and construct boundary | Class and native-context boundary |
|---|---|---|
| 16713569, ATN1/P54259 | Yeast two-hybrid records EBI-950764, EBI-950950 and EBI-951118 retain sufficient regions 1–1185, 372–800 and 785–1190. The available scope is the abstract and curated records, not a recovered full original. | ATN1 is a transcriptional corepressor; its official record includes nucleocytoplasmic shuttling. Neither a native EFEMP1–ATN1 complex nor nuclear EFEMP1 activity is established. |
| 21653829, HOXA1/P49639 | Pooling record EBI-26512440 retains the sufficient-binding region 1–110. Its inclusion in the study's selected mammalian-validation subset is unverified. | HOXA1 is a DNA-binding TF, but its homeobox at 229–288 is outside the recorded fragment. Binding to a segment of that factor does not require contact with its DNA-binding domain or establish a native transcriptional complex. |
| 25910212, MEOX2/P50222 | EBI-12729942, EBI-12702316 and EBI-12715682 are validated, pooling and array records. Two annotate an EFEMP1 residue-345 mutation-with-no-effect feature. Accessible main text is incomplete. | MEOX2 is a DNA-binding TF with a matched reference sequence. Neither native colocalization nor equivalent interaction by all natural EFEMP1 products follows from these records. |
| 26871637, historical MEOX2/Q6FHY5 | EBI-16461465, EBI-16439422 and EBI-16450498 are validated, array and prey-pooling records. Accessible main text is incomplete. | The 303-residue parent equals current MEOX2/P50222 minus Q81 and retains the homeobox. The historical identifier stays intact; this does not authenticate every insert or establish a native complex. |
| 32296183, OTX1/P32242 | HuRI records EBI-22738538, EBI-24262264 and EBI-22732228 include array, validated and prey-pooling assays and annotated GAL4 fusions. | OTX1 has a DNA-binding homeobox and sequence-specific DNA-binding function. Native EFEMP1–OTX1 colocalization remains unestablished. |
| 32296183, HOXA1/P49639 | HuRI records EBI-24694207, EBI-24021820 and EBI-23729189 are distinct from the 21653829 fragment record; the 1–110 interval is not assigned to these HuRI inserts. | The matched parent is a DNA-binding TF. The assay does not establish EFEMP1 transcriptional activity or a native transcriptional complex. |
| 32296183, CLPP/Q16740 | EBI-24652659, EBI-23950645 and EBI-25163909 are validated, array and prey-pooling records. The 277-residue parent retains S153/H178; the mitochondrial transit peptide spans 1–56. | CLPP is a characterized peptidase. The parent checksum does not establish mature mitochondrial processing or a native encounter, and no EFEMP1 substrate cleavage or protease activity is inferred. |
| 32296183, historical MEOX2/Q6FHY5 | EBI-23165847, EBI-24501184 and EBI-22716319 are array, validated and prey-pooling records. | The precise Q81-deletion mapping supports the TF class. Native colocalization and equivalent behavior of every EFEMP1 product remain unproved. |
| 32296183, NR4A3/Q92570 | EBI-23109166, EBI-24463659 and EBI-24977583 are array, validated and prey-pooling records with a matched nuclear-receptor parent. | NR4A3 is a sequence-specific transcription factor. Binding to it does not give EFEMP1 receptor or transcription-factor activity, and native interaction remains unestablished. |
These observations separate confidence in a recorded assay pair from confidence in its partner class and from unresolved native biology. EFEMP1's secretory route and the partners' predominant locations matter to interpretation, but do not demonstrate universal impossibility of contact or a false result in these specific assays. No endogenous human co-immunoprecipitation, native colocalization, mature extracellular folding or natural-product equivalence is asserted for these pairs.
For the three PMID:25406291 IEP assertions, the exact-head committed cache, lines 49–63 supplies a concise cross-check of the study's target panel. This abstract-only file has SHA256 bc2a5090b6e6d300b6c4a9ea71e3e847a3a6777f59691c29978e1ef2ffaed37f; the Methods summary is at lines 49–51. Its machine metadata remains unchanged.
The full original article, PMC4263138 (DOI:10.1167/iovs.14-15453) was read separately. The decisive sections are Methods/Tissue Specimens, Methods/Immunofluorescence, Methods/Microscopy, and Results/Specificity of the Fibrillin Antibodies. The inspected full-text extraction has SHA256 48149b57b0dd38abc48d05510513dff68087bbcfefe6832c3af68f2e518d63f5: Methods lines 109–121, the antibody-panel paragraph at 117, and the specificity-control paragraph at 127. Those original sections identify fibrillin-1, fibrillin-2 and fibrillin-3, MAGP1 and tissue markers, not EFEMP1/fibulin-3. They establish the citation mismatch underlying the existing removals. This full-text access is distinct from the shorter committed cache; no article text is duplicated here and no new quotation is added.