This review evaluates all 45 original annotations from the genuine isolated fetch, preserving their complete source objects and all nine alternative-product objects. The 23 original reference identifiers and titles remain unchanged; a source-evidence index is added as provenance. No NEW annotation is proposed. The scientific decisions are 23 ACCEPT, 19 KEEP_AS_NON_CORE, one MODIFY, one UNDECIDED and one MARK_AS_OVER_ANNOTATED. The derived status is determined by the ordinary validator/status workflow.
COL11A2 is HGNC:2187, UniProt P13942. Original GOA, UniProt, PANTHER, Reactome and publication outputs are preserved; externally recovered full articles are distinguished from the unchanged selected caches.
The genuine default Falcon research attempt timed out after 90 seconds. The configured Perplexity-lite fallback returned HTTP 401 with an explicit insufficient-quota error. The terminal process and error receipts are retained. No provider-named report was fabricated; these manual notes and the actual primary readings supply the review.
Full-length alpha-2(XI) contributes to the collagen-XI heterotrimer and extracellular fibril architecture. The one core uses GO:0030020, collagen type XI trimer GO:0005592, extracellular matrix GO:0031012 and collagen fibril organization GO:0030199. The official GO:0030020 comment explicitly permits extracellular matrix collagens; it does not require asserting that an isolated human alpha-2 chain underwent a force measurement. The broad structural row 3 is refined to this existing specific class rather than creating another core or redundant NEW.
The complete selected human family Methods, Results and Discussion of PMID:10677296(https://pmc.ncbi.nlm.nih.gov/articles/PMC1288089/) establish OSMED variants, RNA and clinical phenotypes. Type-XI composition is described in the background, and a mouse observation in Discussion is a cited personal communication, not a new human reconstitution experiment. The source does not demonstrate adaptor bridging, so NAS row 34 is marked over-annotated for that mechanistic specificity while preserving the established matrix biology.
The complete PMID:9188673(https://pubmed.ncbi.nlm.nih.gov/9188673/) abstract describes three human OSMED siblings and actual cartilage ultrastructure with larger, aggregated fibrils. It is not the mouse knockout paper. The PMID:10581026 original author-uploaded article has separately inspected indexed Methods, Results, Discussion and legends. Human DFNA13 families are distinct from the FVB/N mouse insertion model. Mouse auditory testing used eight animals and showed impaired homozygotes; a separate six-animal histology set showed abnormal tectorial fibril organization. Neither the indexed access nor the normal abstract is represented as a complete local PDF or a figure-pixel inspection. Hearing and developmental annotations remain contextual consequences of the structural role.
The bounded independent core consultation agrees with this integrated core and source-scoped adaptor assessment. Its same-role collagen comparator query confirms existing structural/fibril annotation convention; it does not manufacture a new process assertion. No literal P13942/COL11A2 entry occurs in the selected GO-CAM index; this index result is not evidence of biological absence.
The original PMID:17703188 publisher article was recovered as genuine full HTML and its actual Results, Methods, Discussion and embedded legends were read. Human DDR2 residues 26–186 were expressed in Pichia. NMR uses bovine nasal-septum type-II collagen; SPR uses commercial type-II/type-IV substrates with an unresolved human-placenta parenthetical. A GPO-repeat helix is a computational docking model, not an assayed sequence-defined COL11A2 chain.
Both literal positive MINT records mapping P13942 to Q16832 are preserved with every field, including their in-vitro host and database-assigned human participant identities. Those database sequence identities do not establish the composition of the experimental collagen preparation. Row 4 is UNDECIDED at the exact COL11A2 link, and the reference is UNVERIFIED specifically for that link. This does not deny the positive DDR2–type-II result, claim universal noninteraction, or retarget the source assertion. The original SPR text and a literal MINT kinetic note differ numerically; neither source record is rewritten and no kinetic value is used in the authored biological claim. Full-text access resolves what the paper reports, while the exact tested-chain mapping remains unadjudicated. The campaign instruction retains UNDECIDED in that situation; genericity alone is not grounds for removal. The source-evidence index, under ddr2_mapping_clarification, provides the resolved publisher URL, source hash, section locators and short primary excerpts.
The protected main PMID:28344315(https://www.nature.com/articles/leu2017102) abstract is unchanged. Genuine publisher Methods/Results and the original S2/S3 workbooks were read separately. The exact P13942/COL11A2 complete-data row records seven unique peptides and detections in four healthy donors, one newly diagnosed sample and one relapsed sample. S2 applies a different cohort-signature classification, so its disease “nd” field is not used to claim absence from all diseased samples. This supports presence in human bone-marrow ECM-enriched preparations, not a force measurement, a unique producer cell or resolved isoform abundance.
For PMID:28675934(https://pubmed.ncbi.nlm.nih.gov/28675934/), a separate genuine normal full variant and authentic PMC HTML were read while the main cache remains exact. Methods establish ECM enrichment and human tissue proteomics; supplemental links returned HTML download challenges rather than readable workbooks. The target table remains uninspected. Existing structural/location assertions are retained from independently supported collagen identity and target ECM evidence, without falsely marking the unresolved exact citation link as verified.
Reactome target chains reach reference entity 51172, UniProt P13942, for all nine original events. ER-luminal precursor/trimer residence remains non-core; extracellular collagen processing and assembly contexts remain positive. A collagen input to a cleavage event is not the catalyst. Multiple product/entity paths are curated-model representations, not independent experiments. The broad extracellular-region TAS terms deliberately retain the source compartments, including processing-substrate states. Independently supported extracellular-matrix annotations describe matrix-associated product states; they do not require every extracellular precursor or propeptide entity to be retargeted to matrix. This differs from refining the structural molecular-function class to the supported collagen-specific term.
The nine preserved products reconstruct to 1736, 1710, 1715, 1676, 1689, 1650, 1655, 1629 and 290 residues using the actual current UniProt VAR_SEQ edits. Eight independently fetched official FASTAs match. The isoform-2 request failed, and its reconstruction is explicitly based on the canonical sequence plus official edits rather than an invented successful FASTA fetch. The short product and every splice form are not assumed to assemble or function like full-length collagen; the product-specific question remains open.
All 63 extant source-entity/term joins are mapped to current official identifiers, with 348 exact current annotation records. Current experimental and inferred-only records are distinguished in the propagation comments. Historical PAINT nodes are preserved and not reconstructed from current donor counts. Target self-evidence is legitimate descendant grounding. No claim of reading every donor primary paper or finding a historical node-placement defect is made.
The original abstracts of 7859284, 8838804, 9188673, 12673280 and 16033917 were read completely and remain abstract-only access. The target passage in the family review 17876790 remains unavailable; the original NAS trimer/location assertions have independent support. Full-text flags distinguish these limits from external full recovery. No cache is hand-written or upgraded in place. Two short load-bearing primary anchors occur only in YAML, not duplicated in these notes or the source index. The reproducible read ledger, literal tables and cumulative quote audit are retained with the finite packet.