This manual review preserves the genuine 53-row source seed and all selected publication, UniProt, GOA and PANTHER source bytes. The normal Falcon request failed with a service error and its configured Perplexity fallback failed authentication. Those receipts are retained; no provider text was invented. The isolated runtime uses Python 3.12 and the exact e2057 tooling, including linkml-reference-validator 0.3.1rc1. Its own environment is independent of the publisher.
CTSK is human HGNC:2536/P43235. The current official precursor is 329 residues, with signal peptide 1–15, propeptide 16–114 and mature catalytic chain 115–329 (215 residues). The sequence independently agrees with the official FASTA and family alignment. The mature 1MEM chain is exact and contains the covalent catalytic-cysteine inhibitor geometry. Processing does not create a second splice product. Historical cathepsin O/O2 nomenclature in the original CTSK papers is distinguished from present-day CTSO/P43234.
The main biological unit is cysteine endoproteolysis with prominent collagen and extracellular matrix catabolism. The human recombinant digestions in PMID:19834056(https://pmc.ncbi.nlm.nih.gov/articles/PMC2774069/) distinguish performed proteolysis from mouse tissue phenotypes and demonstrate glycosaminoglycan-dependent modulation. PMID:22952693(https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0043494) contains actual CTSK substrate and secretion experiments despite focusing on cathepsin S. Its authentic supplementary Figure S2 was visually read: CTSK cleaves collagen IV, fibronectin and perlecan with E-64 controls. The SPR and ELISA experiments primarily test nidogen partners and do not establish a CTSK equilibrium binding constant. Human enzyme and mouse perlecan identities remain explicit.
The positive SERPINB13 and CST7 interactions are kept as non-core binding under the CLINGEN_MENDELIAN Curation instructions, which override generic-binding deletion guidance. These are inhibitor–enzyme interactions; they do not confer inhibitor activity on CTSK. The cystatin F original was personally read through genuine indexed publisher Results, Table 1, Discussion and Methods, while its normal cache remains abstract-only. Its U937 microscopy does not show CTSK colocalization with cystatin F. The complete SERPINB13 abstract directly reports inhibitory complexes; no full-paper access is claimed.
ROOT supplied a bounded independent thyroid/mitophagy consultation, fully read and reproduced by the owner. PMID:11082042(https://doi.org/10.1242/jcs.113.24.4487) remains abstract/first-page only; porcine cloning is explicit, but the human tissue and reagent details of every assay are unresolved. Established human annotations are retained with curator deference, not rejected from that partial view. The reported thyroglobulin cleavage is performed proteolytic work. In the separately inspected mouse primary, explicit apical CTSK staining is in Ctsb-deficient thyroids; Ctsk/Ctsl double deficiency lowers thyroxine while Ctsk single deficiency is compensated. The thyroid and surface contexts remain non-core.
The actual original mitophagy Table 1 identifies human CTSK. ROOT read the indexed Results and human RNAi Methods, which are absent from the selected cache despite its full-text flag. The contextual screen result is retained without transferring SREBF1-specific lipogenesis, PINK1 stabilization or rescue mechanisms to CTSK. No target-specific supplement statistics or rescue was recovered.
The genuine PANTHER v19 topology has 4,609 nodes. CTSK lies below the four positive PTN000274640 IBDs, and no relevant actual GAF loss record lies on its 22-node path. Its alignment exactly matches the 329-residue official protein. The tree leaf nevertheless decorates catalytic/proteolytic slim terms as NOT; this contradictory display is preserved as an unresolved metadata question, not treated as catalytic loss. Current extracellular IBD data add two zebrafish identifiers to the unchanged GOA source list. Complete donor queries and official/tree aliases are retained in the audit workspace; overlapping zebrafish accessions are not counted as independent experiments. Target self-evidence is legitimate.
All six current ARBA rules were inspected. Five match current target condition sets; ARBA00028064 alone does not. Branch 0 of both ARBA00086247 and ARBA00090216 matches FunFam 3.90.70.10:FF:000006 together with Primates (9443). Those release/snapshot differences are distinguished from the unchanged historical GOA provenance and independently established biological activity. The nine seeded Reactome roots have actual positive P43235 paths through 133 retrieved objects. The three serine-type MF transfers are removed because CTSK uses a cysteine nucleophile; extracellular localization and valid matrix substrate contexts remain. Other endostatin-set catalyst labels raise the same model-level chemistry concern without generating new target assertions. The production GO-CAM activity index has no literal CTSK hit at the research boundary; that does not establish absence from every model input.
The twenty current IntAct records contain none of the four seeded IPI pair/publication combinations. This is not negative experimental evidence: the original substrate/inhibitor experiments and stated curator deference supply the relevant positive support. No extra functions were proposed from unrelated current interaction records.
The owner previously authored the immutable three-row catalytic peer; that reuse is explicitly attributed and is not represented as another independent review. ROOT separately read and reproduced it. Full primary read boundaries and failures remain in the saved READ ledger. The compact source guide provides public and repository locators without dumping raw data or repeating every review decision. All verbatim excerpts are short cache-exact strings, with aggregate counts including repeated placements no greater than 25 words per publication. No additional primary quotations are embedded in these notes.
The genuine normal publication cache for PMID:12782676(https://doi.org/10.1172/JCI15990) was added after a source-path check. It is abstract-only; the independently inspected original JCI Methods, Results and captions remain the basis for the mouse thyroid distinctions above. The formal bibliography and rows 29, 35 and 44 now cite that source explicitly. No scientific action or quotation changed.
The original indexed Results, human RNAi Methods and Table 1 of PMID:24912190(https://pmc.ncbi.nlm.nih.gov/articles/PMC4060696/) were reread. CTSK is one of 20 human hits, selected in triplicate YFP-Parkin HeLa assays using both early Parkin translocation and later mitochondrial clearance. The existing HMP annotation remains KEEP_AS_NON_CORE. This is a contextual experimental inference, without a demonstrated CTSK mitochondrial substrate or selective mitophagy mechanism. The cache alone is incomplete; its omission of the table is not absence of the original target result. Neither the SREBF1 follow-up mechanism nor a proposed nonspecific lysosomal-flux explanation is assigned to CTSK.
Collagen binding is now KEEP_AS_NON_CORE, consistently with fibronectin and perlecan substrate engagement. Actual supplementary Figure S2 of PMID:22952693(https://doi.org/10.1371/journal.pone.0043494) shows inhibitor-sensitive human CTSK cleavage of human collagen IV. That supports the curated contact in a catalytic context, without a separate adhesive or affinity-measured binding activity. The single cysteine-endopeptidase core retains collagen catabolism and extracellular-matrix disassembly.
PMID:11082042(https://doi.org/10.1242/jcs.113.24.4487) is now an explicit supporting pointer for the specialized external-surface localization. Its abstract reports an active apical surface pool, while complete assay-species and genotype details remain unresolved. The distinct mouse PMID:12782676(https://www.jci.org/articles/view/15990) observation is specifically apical staining in Ctsb-deficient thyrocytes. Neither result is generalized to every tissue.
Bone resorption (GO:0045453) is part of bone remodeling (GO:0046849). The former includes osteoclast degradation of bone material. PMID:19834056(https://pmc.ncbi.nlm.nih.gov/articles/PMC2774069/) demonstrates human recombinant CTSK collagen and mouse bone-powder digestion, alongside mouse osteoclast/dentine experiments. Thus CTSK performs a matrix-degrading step. However, current human bone-resorption descendant queries were empty for CTSK and the role-relevant matrix proteases MMP9, MMP13 and MMP14; positive query controls succeeded. These comparators are not identical osteoclast contexts. No NEW process or redundant core list is added; a focused question records the annotation-scope issue.
The two broad proteolysis and four protein-catabolism assertions remain ACCEPT: the source evidence covers multiple protein substrates, including noncollagen matrix proteins and thyroglobulin. Their breadth is therefore useful at the source level, while the core is selective. Narrowing broad catalytic MF assertions to cysteine endopeptidase does not require replacing all broad BP sources with collagen-only terms. The compact source guide now supplies public Reactome ContentService locators for the mechanistically incompatible shared serine-catalyst path; genuine machine caches and source tuples remain unchanged.