CDKN2A notes

2026-10-10 — ClinGen Mendelian review

This new review uses authenticated main ffaf00ab93d40e5fe2e71b283ac42f10ffb22d2f for eligibility and existing-source bytes. The frozen catalog listed CDKN2A as unchecked; complete changed-path checks of 169 open PRs and HGNC aliases found no overlapping owner. No earlier CDKN2A notes existed at that main revision. The normal fetch produced 79 original P42771 assertions, six alternative products and 42 initial references. Every source annotation, qualifier, partner, product and fetched title is preserved. No NEW assertion is introduced.

The normal default Falcon attempt timed out at its bounded limit; the genuine perplexity-lite fallback returned an insufficient-quota error. Neither produced a report. This journal and the accompanying manual evidence are explicitly manual research. Authentic normal source caches are preserved, with existing-main versions selected byte-for-byte and fresh variants archived separately.

Biological synthesis

p16INK4A is an INK4 inhibitor of CDK4/6, not a kinase or cyclin-binding enzyme. Its binding limits cyclin-D-dependent phosphorylation of RB and restrains G1/S progression and proliferation; it contributes to stable arrest in senescence contexts [PMID:8259215, PMID:9751050, PMID:10208428, PMID:16243918]. The core includes nuclear and cytosolic regulatory pools. Observed subcellular localization, PAINT activity-in-compartment assertions, and the curated cytosolic INK4–CDK reaction are distinguished rather than treating staining alone as proof of activity.

The unrelated p14ARF reading-frame product is explicitly mapped to Q8N726-1 in functional_isoforms. Its MDM2/p53 mechanism and bounded N32-fusion ligase inhibition are separate from the P42771 core [PMID:9724636, PMID:35944929]. Full primary construct and knockout checks resolve five source/product errors: one ARF/spinophilin interaction and four transfers of ARF-dependent apoptosis from the mouse p16 accession. These are not generic-binding policy removals and do not negate the valid ARF biology. No untested function is transferred to smARF, p12 or every splice product.

Annotation decisions and standing authority

Final actions are 32 ACCEPT, 24 MODIFY, 15 KEEP_AS_NON_CORE, five REMOVE, two UNDECIDED and one MARK_AS_OVER_ANNOTATED. The 24 refinements comprise 23 source-specific CDK4/6 binding assertions plus one population-proliferation correction. Distinct source rows are preserved even when their replacement terms coincide.

The immutable project instruction is the authority for supported generic binding retention. It says to retain a supported, biologically correct GO:0005515 annotation as KEEP_AS_NON_CORE when no evidence-backed, more specific replacement has been established. It also requires evidence-backed MODIFY and unresolved UNDECIDED. This explicit user instruction overrides the annotation-reviewer skill's general informational-exclusion recommendation; it does not verify an uninspected pair or establish a narrower activity from a partner name alone.

Protein-kinase binding refinements rely on verified CDK4/6 associations. Their screen methods do not individually establish inhibition; p16 inhibitor activity has separate biochemical support. Other source-linked associations remain non-core at their measured resolution. IntAct, BioGRID and GOA may share curation provenance and are not counted as independent replication.

The negative cell-growth row measured serial viable-cell counts, DNA synthesis, colonies and cell-cycle profiles; it is refined to negative regulation of cell population proliferation, with early cytostasis separated from later death PMID:10208428. The Ras IEP assertion is over-annotated as a relay role: induced p16 and delayed arrest do not show that p16 relays the Ras signal PMID:9054499. Non-core matrix adhesion and NF-kappaB regulation retain the exact assay contexts rather than becoming additional core functions.

Two explicit uncertainties remain. The RNA-capture dataset has a positive classIII CDKN2A entry but does not resolve p16-specific peptides versus ARF. The senescence-focus source supports formation/arrest but its inaccessible full experiment leaves p16 residence in the focus unadjudicated. These are evidence-bound uncertainties, not negative functional claims.

PAINT assertions retain their phylogenetic meaning. Neither a short donor list nor target-self support is treated as weakness/circularity. Exact ancestral node placement was not reconstructed; independently established target function and the absence of a specific loss claim support retention. ARBA and location-mapping rows likewise retain provenance without claiming a newly reconstructed whole rule.

Source access

The manual primary evidence, interaction extract, screen extract and product/donor extract provide source URLs, exact record fields and bounded assay/construct interpretation. Existing abstract-only caches were not rewritten to imitate full text. Reference review notes state both the cache status and actual external access; no full-text-unavailable flag contradicts a successful external reading.

No new quotations appear in this notes journal. Short primary anchors appear once each in YAML or the manual-evidence file. The authoring audit counts repeated excerpts across all authored files against a 25-word-per-source limit; this is an authoring constraint, not an asserted repository validator rule.

Reference reading ledger

2026-10-10 accession and source follow-up

This entry supersedes the earlier functional-class placement and four shared interaction reasons; the preceding journal is preserved verbatim. The actual PR-head alternative-ORF rule requires the unrelated ARF protein to have its own accession folder. Both functional_isoforms entries are removed from this P42771 review. The primary p16 inhibitor core remains. UniProt's six alternative-product descriptions, including external Q8N726 entries, remain unchanged; the source listing is not declared erroneous. The separate Q8N726 review substantively reviews its own 72 seeded annotations. It does not synthesize ARF source tags on p16 rows or merely relocate the old assertions.

The five p16 product-mismatch actions remain REMOVE after the prior full primary construct/genotype assessment. The valid ARF experiments remain described in the prior evidence and accession-specific discussion, but are not manufactured into new transferred annotations. The full-length human ARF–spinophilin assay and mouse ARF apoptosis findings do not establish p16 activities.

RNA row 2 remains UNDECIDED. H1 is genuinely named the label-swap experiment; the raw table is heavy/light. Light is crosslinked in L1/L2 and heavy in H1, so normalized crosslinked/control log2 enrichment is +3.31 and +1.39, respectively, yielding mean +2.35. Both measured observations enrich, only L1 exceeds threefold, and L2 is missing. Neither class III nor lack of a canonical RNA-binding domain proves an erroneous interaction. The inspected gene-symbol row and protein-group Methods still do not resolve P42771 versus Q8N726 peptides.

Rows 21–24 now have individual partner reasons. Authentic PMID:21988832 supplementary Table S1 identifies the bait as p16; Table S2 and eight exact positive IntAct records support engineered DEAF1, HNRNPU, TDRD7 and CRELD2 associations. The first three remain non-core with compatible intracellular context. CRELD2 becomes UNDECIDED: its signal peptide/ER routing raises a native encounter question that the fragment/tag/lysis assays do not settle. Positive assays are acknowledged; no artifact or impossible interaction is asserted. The compact source extract preserves literal pair, method, host and source records. These database records describe the original experiments, not independent replication.

The standing project instruction continues to govern supported generic association. All 79 annotation source objects, six products and selected existing caches are unchanged. No new quotations are authored in this append; historical quotations above are inherited. No NEW annotation is added.

2026-10-10 — second review follow-up

The RNA-binding question now asks both which CDKN2A product supplied the captured peptides and whether independent direct RNA-binding assays on defined p16 establish the interaction and its specificity. Product resolution alone would not settle the functional interpretation. The existing UNDECIDED action and source-specific RNA evidence assessment are unchanged, as are all 79 source assertions, products and other decisions.