The dated entries below are a historical journal. Later entries supersede earlier decisions where explicitly stated; the current annotation judgments are in the YAML. During the October audit, repeated source quotations were reduced to citations or paraphrases while retaining the dates and scientific history.
PITA context: CDH23 corresponds to PITA5 / pituitary adenoma 5, multiple types. UniProt notes that PITA5 includes GH-, PRL-, ACTH-, TSH-secreting and plurihormonal tumors and that familial transmission is consistent with autosomal dominant inheritance with reduced penetrance [file:human/CDH23/CDH23-uniprot.txt].
Deep research status: just deep-research-falcon human CDH23 --fallback perplexity-lite timed out on Falcon after 600 seconds, and the fallback failed with a Perplexity quota error. I proceeded using fetched UniProt, GOA, and cached publications.
Functional summary: CDH23 is a cadherin-family adhesion protein whose best-supported normal function is in stereocilia/hair-bundle architecture. UniProt describes calcium-dependent cadherin adhesion and a role for CDH23 in hair-bundle architecture [file:human/CDH23/CDH23-uniprot.txt]. Experimental evidence shows cadherin 23 is present in growing stereocilia and binds harmonin PMID:12485990, and CDH23/PCDH15 maintain normal stereocilia bundle organization PMID:15537665.
Annotation decisions: I accepted cadherin adhesion, calcium binding, membrane/stereocilium localization, and sensory phenotypes as supported, with sensory phenotypes kept non-core where they are phenotype-level consequences. I removed the calcium ion transport annotation because the cited PMCA2 paper describes stereociliary calcium entry/export biology rather than CDH23 transporter activity PMID:17234811.
All 25 source rows, core functions and references were assessed; source fields were preserved. PTHR24026 IBDs were traced to PTN008601603 (adhesion, migration, beta-catenin binding, catenin complex, cadherin binding) and PTN004649584 (neuron-projection development). No matching repository or global OpenScientist report was found. Root received neutral questions for catenin interactions/complex, neuronal process and calcium processes; those require coordinated adjudication. No report was independently launched.
The old heterophilic-only/hair-cell-only exclusions are contradicted by target-cell studies. PMID:22413011 reports that antibody inhibition and CDH23-directed RNAi reduce human MCF-7 cell adhesion. Homotypic cells alone do not prove a homophilic molecular interface, but PMID:31729176 directly studies mouse NP_075859.2 EC1-2 trans-homodimers and a conserved binding interface, with human A549-cell adhesion tests. This ortholog construct scope is recorded explicitly. PMID:30747484 links CDH23 silencing to less aggregation and greater migration in human cultured cells. These support restoring homophilic adhesion and retaining migration as secondary.
Beta-catenin remains a separate mechanistic question. PMID:12485990 reports absent consensus R1/R2 motifs and unpublished lack of catenin in growing hair bundles. PMID:22413011 and PMID:30747484 show beta-catenin-positive CDH23 cell contacts, but colocalization is not direct binding or proof of catenin-complex membership. These two IBAs are UNDECIDED; neuronal development is retained as non-core pending further mechanistic assessment because the inherited role is not refuted by sensory specialization.
The full PMID:17234811 results assay PMCA2 variants for calcium handling and examine CDH23 genetic variation in the deafness family. The paper leaves the relationship between the CDH23-containing complex and PMCA2 function unresolved. The old claim that only a pump could participate in calcium transport was invalid; the two specific calcium processes are now UNDECIDED pending assessment of mechanical channel coupling and the original IMP interpretation.
PDZ2 binding is directly tested in PMID:12485990, supporting modification of generic protein binding to GO:0030165. PMID:22879593 explicitly describes exon68-dependent self-association and the C-terminal PDZ motif. Its full text remained unavailable after fetch and PMC browser attempts; only abstract-supported details are used. Human USH1 genetics (PMID:16679490) supports hearing, balance and retinal roles. The description and cores now distinguish extracellular cadherin interactions from intracellular PDZ-domain binding.
Read the complete catenin/neuronal/calcium report and checked its decisive claims against the already reviewed primaries. Its recurring statement that PMID:30747484 contains “zero occurrences” of catenin is false. The full Results and Discussion describe CDH23 at beta-catenin-positive contacts, including endogenous protein in A549 cells and plasmid-expressing HEK293T cells. The same Discussion identifies a possible MAGI-1 bridge and reports no significant canonical β-catenin-pathway change after silencing (Fig. S8). None proves direct binding or a defined catenin complex; retain both UNDECIDED. The prior notes' attribution of colocalization to this paper was correct.
The report's missing classical motif/harmonin evidence remains a real mechanistic concern, but an alternate native linkage is not a universal negative binding assay. Its neuronal exclusion depends on missing target experiments and limited zebrafish retinal morphology/optokinetic assays, not a demonstrated loss of the human inherited developmental capacity; retain non-core. For calcium processes, the full PMID:17234811 PMCA2 assays and explicit open question remain decisive limitations; a structural component can do work in transport without being the pump. Both IMP process claims remain UNDECIDED. Mark the report DISPUTED, preserve its substantive evidence and require human source/mechanism follow-up rather than another duplicate report.
Preserve superseded donor-specific analysis where present, replace rebutted report passages with actual supporting evidence, and clarify location or process scope. PAINT rows support inherited assertions as phylogenetic judgments, not direct target experiments.
Reassessed all 25 existing source assertions from main 5c3a43a48ef70fac4f5438c17cd2989515e50865, including the eleven product records. Fresh normal intake independently seeded 25 rows; its bytes are archived separately. The campaign candidate preserves the exact-main annotation objects outside review, product metadata, UniProt/GOA bytes, publication caches and the existing OpenScientist report. It changes no original action and adds no annotation. The earlier COMPLETE status is a document status, not a claim that this campaign audit has passed peer review and CI.
The standing ClinGen project instruction retains correct supported generic binding as non-core when a narrower MF is not established. Here the two generic interaction assertions already have a source-supported PDZ-domain-binding refinement, retained after direct interface inspection; no generic binding is removed for being uninformative.
The extracellular core remains GO:0045296 cadherin binding. This legitimately includes the CDH23-PCDH15 sensory contact between stereocilia of one hair cell, as well as separately demonstrated adhesion between cells. GO:0098641 has the additional cell-cell process restriction and would not describe the principal sensory contact. The core now explicitly distinguishes these contexts and names the existing accepted hearing, balance and cell-cell-adhesion processes. Human MCF-7 antibody/RNAi adhesion assays in PMID:22413011 Results/Fig.3 support the latter independently of tip links. PMID:31729176 Methods uses mouse NP_075859.2 EC1-2 constructs; A549 adhesion to immobilized protein and HEK293 aggregation after EC1-2 transfection provide complementary cellular assays. These are not purified-human ectodomain experiments. Endogenous CDH23 silencing in PMID:30747484 supports the migration decision, while its gift-plasmid species/accession was not established by the inspected Methods.
A newly accessible author manuscript for PMID:22879593 resolves an inherited access limitation without replacing the abstract-only main cache. Protein Preparation specifies human CDH23 cytoplasmic tails with or without exon68 and human harmonin-a. Results/Fig.4 shows that the +exon68 tail forms a stable dimer, whereas the isolated GB1-exon68 construct aggregates; flanking sequences stabilize the tail dimer. The self-binding reason now reflects that distinction. Results/Figs.5-6 distinguish the C-terminal PDZ-binding motif from two NTD-binding sites and show their contributions to multivalent assemblies. The experiments measure recombinant binding/assembly; stable anchorage under force in living hair cells is an interpretation, not a direct assay here. PMID:12485990 independently tests a human CDH23 tail against isolated mouse harmonin PDZ2 with PDZ1/PDZ3 controls (Fig.6; Expression constructs). Harmonin's actin-binding activity is not assigned to CDH23. Selected Methods/Results and figure legends were read; figure pixels and all supplements were not separately inspected.
The four UNDECIDED decisions remain evidence-specific. For beta-catenin binding and catenin-complex membership, motif divergence and an alternative harmonin interface raise a mechanistic concern, while native human colocalization in PMID:22413011 and PMID:30747484 does not prove either a binary interface or a defined complex. MAGI-1 is a proposed bridge in the latter Discussion. The preserved OpenScientist report's claim that PMID:30747484 lacks catenin text is demonstrably false. For calcium transport and cytosolic calcium regulation, full PMID:17234811 Results/Discussion/Methods separate the CDH23 T1999S family modifier hypothesis (also present in healthy relatives) from PMCA2-specific CHO-cell calcium-clearance experiments and PMCA2-mutant/knockout mouse imaging/current measurements. These experiments do not independently assay CDH23 calcium handling. A mechanical linker can contribute to channel gating without being a pump, so the missing pump activity is not used to reject the existing process assertions.
The existing-main PAINT table was read directly: PTN008601603 carries the cadherin/catenin/adhesion/migration IBDs and PTN004649584 the neuron-projection IBD. Those are phylogenetic judgments; neither donor count nor lack of a target experiment is used as a reason to reject them. The tree/MSA was not newly reconstructed, and the uncertain catenin placement remains unresolved. Clinical source PMID:16679490 identifies CDH23-mutant USH1 families and supports existing auditory, balance and retinal assertions at their stated genetic scope; it does not establish a phototransduction enzyme mechanism.
Normal provider research was genuinely attempted. An initial call started before UniProt completion and failed template expansion; the post-intake retry timed out on Falcon at its 90-second limit, and the explicit Perplexity-lite fallback returned a quota error. The single orphan Falcon child retaining the output pipe was terminated after the launcher exited, with a separate receipt. No new provider report was produced or synthesized. Manual evidence above and preserved primary caches support the audit.
The audit condensed repeated source quotations into short anchors and reference-specific paraphrases. The subsequent citation follow-up below restores distinct evidence anchors that the initial compaction omitted; the original scientific assertions and provider output remain preserved. The publication-cache refresh described below is genuine tool output.
The review accepted the biological decisions and requested more checkable evidence. The earlier compaction reduced 46 YAML quotations to five; this follow-up restores short, distinct anchors for MCF-7 RNAi adhesion (PMID:22413011, Fig.3), the missing classical beta-catenin motifs (PMID:12485990, Discussion), beta-catenin colocalization (PMID:30747484, Results/Fig.1), the exact neuron-development PAINT IBD record, and the PDZ-motif/self-association distinctions in PMID:22879593. Repeated copies of the same excerpt are not needed on every membrane or adhesion assertion. GO:0045296 now also records PMID:22413011 in supported_by, matching its existing reason. All 25 original source objects, all decisions, eleven products, description and biological core assignments remain unchanged. Historical duplicate citation tokens above were collapsed; dated scientific conclusions were not rewritten in this follow-up.
The distinction underlying the catenin decisions remains checkable: PMID:30747484 locates CDH23 at “cell–cell boundaries”; the longer YAML anchor identifies beta-catenin colocalization, which does not itself establish direct binding or defined complex membership. The calcium-process uncertainty is similarly source-specific: PMID:17234811 says the complex's relationship to PMCA2 “is an open question”. These short notes anchors complement the separate YAML excerpts.
The repository's supported refresh-publications --dir <isolated-publications> --count 1 command recovered the complete PMC3460447 primary text through its HTML fallback for PMID:22879593, DOI:10.1074/jbc.M112.378505. The refreshed publications/PMID_22879593.md is unedited tool output, SHA-256 d58e89329fed937e8e84dcf13f37a57c0d0de2b9967a0f895e04477c6b0e97cd. Protein preparation and Results/Fig.4 document the human constructs, analytical-ultracentrifugation distinction between +exon68 and -exon68 tails, and the separate aggregation of isolated GB1-exon68. Results/Figs.5-6 describe the multivalent harmonin interfaces. Selected full text and legends were checked against the earlier primary manuscript; this is not a claim to have independently inspected every figure pixel or supplement.
The previously read author-laboratory manuscript remains an authentic primary source: DOI and PMCID above identify the paper independently of its hosting path, and its PDF SHA-256 is a901549e74e1afc65b9c5068da1c1e90eab93c7e1241055d93f3013ff21d74fe. The prior abstract-only cache was a cache-coverage limitation, not lack of full-text access by the reviewer. Its reference assessment therefore remains VERIFIED, without a false full_text_unavailable flag. The two other papers flagged unavailable retain their separate access limitations.