The canonical seed contains 14 annotations, eight source terms, four PMID references and four GO reference entries. HGNC:21316 / UniProt Q6UB99 identifies human ANKRD11, also ANCO1. The source seed, UniProt and GOA are preserved; the ordinary remote source pipeline supplied them unchanged after the original local DNS failure. This is manual research, not a provider-generated report. The ordinary Falcon wrapper first failed at the protected default uv tool location; a configured attempt with the installed matching client reached Falcon authentication/network handling and the Perplexity fallback, both of which failed. No provider report was produced.
The four normally fetched source publication records were read in full at their actual cached scope: PMID15184363, PMID25413698 and PMID25556659 are abstract-only; PMID21782149 has an extracted full main article. External primary sources were inspected separately, without altering those caches.
PMID:15184363(https://pubmed.ncbi.nlm.nih.gov/15184363/), Zhang et al. 2004, identifies ANCO-1 as a p160-coactivator-interacting nuclear protein. The abstract describes nuclear foci, HDAC association and repression of ligand-dependent nuclear-receptor reporters. It proposes recruitment of HDACs; the abstract alone does not establish every step of that recruitment model. The original full article remains uninspected.
PMID:17521611(https://pmc.ncbi.nlm.nih.gov/articles/PMC1950474/), Zhang et al. 2007, was inspected through the original main Methods/Results and figure legends. Gal4 fusions mapped activation and repression domains; full-length ANCO-1 repressed reporter transcription, while knockdown increased progesterone-receptor-mediated transcription in HeLa cells. COS-7 and yeast were expression/assay hosts, not grounds to assign the human protein a different species. These experiments do not establish native sequence-specific DNA binding: the Gal4 DNA-binding domain supplied targeting. The proposed simultaneous p160/HDAC recruitment remains a model rather than a directly visualized complex.
PMID:18840648(https://pubmed.ncbi.nlm.nih.gov/18840648/), Neilsen et al. 2008, was verified against PubMed and read at abstract scope. Human breast cancer cell experiments identify ANKRD11 as a p53-interacting coactivator, link it to p53 cofactors P/CAF and hADA3, and show enhanced p53 acetylation and CDKN1A expression. Acetylation is performed by partner enzymes; ANKRD11 is not thereby an acetyltransferase. Native p53 and mutant-p53 experimental contexts must be distinguished. The full original paper has not yet been inspected.
PMID:25556659(https://lab.research.sickkids.ca/wp-content/uploads/sites/68/2018/02/Ankrd11-is-a-Chromatin.pdf), Gallagher et al. 2015, was additionally inspected in the author-lab copy of the primary Dev Cell article. Figure 2 and the human-precursor Results show nuclear ANKRD11 in human ESC-derived cortical precursors and neurons. Human knockdown reduced precursor proliferation and neurogenesis. Mouse cortical-precursor experiments support chromatin association, changed target-gene expression and histone acetylation, HDAC3 colocalization/proximity and rescue by HDAC3 expression or acetyltransferase inhibition. Proximity ligation within 40 nm is not purified binary binding, and the mouse mechanistic experiments are not mislabeled direct human experiments. Supplemental files were not inspected.
PMID:39847329(https://pmc.ncbi.nlm.nih.gov/articles/PMC11789155/), Liu et al. 2025, was verified against PubMed and the primary main Results read. Purified human ANKRD11 peptide 342–378 binds STAG2–RAD21 with approximately 0.1 micromolar affinity; the 3.2 angstrom structure and biochemical competition show overlapping binding with CTCF. Full-length human ANKRD11 pull-down in HeLa cells and the Y347A binding-disrupting mutation connect this site to cohesin association. Endogenous tagged mouse ESCs and knock-in mice provide distinct gene-expression/developmental evidence. The phenotype does not assign ANKRD11 cohesin ATPase activity, DNA-loop extrusion or sister-chromatid cohesion; neither is physical association sufficient to declare ANKRD11 an obligate cohesin subunit. Supplemental experiments remain uninspected.
PMID:21782149(https://pmc.ncbi.nlm.nih.gov/articles/PMC3155157/), Sirmaci et al. 2011, reports human ANKRD11 variants with dental, skeletal and craniofacial phenotypes. This supports the existing developmental perturbation phenotypes, without demonstrating a tissue-specific molecular step for each phenotype. The GFP-ANKRD11 localization experiment used mouse postnatal cortical cultures; the primary paper reports predominantly nuclear fluorescence with a small cytoplasmic pool. Depolarization redistributed nuclear signal into inclusions. The host species must not be conflated with the expressed construct, whose exact sequence species was not stated in the inspected main Methods.
PMID:25413698(https://pubmed.ncbi.nlm.nih.gov/25413698/), Walz et al. 2015, reports variant-dependent protein accumulation and C-terminal degradation control in the abstract. The full paper has not been recovered; absence of a localization sentence from the abstract is not a reason to reject its curated nuclear annotation. No ubiquitin-ligase or proteasome catalytic activity is inferred for the protein undergoing degradation.
Human Protein Atlas, ANKRD11, inspected 2026-09-27, reports enhanced nucleoplasm as main location and enhanced cytosol as an additional location. CAB019288 labels both in A-431, U-251MG and U2OS; HPA041593 labels nucleoplasm in A-431/U2OS and both in SiHa; HPA049470 labels both in A-431/U-251MG. The numeric cell-line column is RNA expression (nTPM), not antibody signal intensity. The direct database summary is: “Mainly localized to the nucleoplasm. In addition localized to the cytosol.” The cytosolic pool is a real localization observation, not yet a separately characterized core molecular role.
GO:0003712, transcription coregulator activity, is an MF term for modulating transcription through interactions with DNA-bound transcription factors or their associated coregulators. Its corepressor and coactivator children describe distinct contexts; an unsigned coregulator description can summarize the two without inventing DNA-binding specificity or catalytic histone modification. Existing IBA nucleoplasm evidence includes the target itself; this is legitimate PAINT descendant evidence, not circularity. No ANKRD11/Q6UB99 match was found in the local GO-CAM index.
Review decisions and independent annotation consultation are still pending.
All three additional normal PMID records were recovered in Source28 and imported exactly without overwriting existing caches. PMID:17521611 includes the main article, PMID:18840648 remains abstract-only, and PMID:39847329 contains a repeated main-text extraction with supplements uninspected. Actual reading scopes and hashes are recorded in tmp/ANKRD11-initial/source28-actual-source-assessment.json. All 14 original annotations and alternative products remain intact; seven annotations are accepted, seven retained as non-core, and one transcription-coregulator molecular function is proposed. No new biological-process row is added. Independent authored-draft consultation prompted removal of qualifier-based reasoning from three developmental rationales; decisions depend on the experimental phenotype evidence. Final validation and publication follow this source closure.
Final independent source-closure consultation passed on the completed 15-row review. Schema/reference validation, rendering and history validation passed. The sole intentional annotation advisory concerns GO:0006357 in the core synthesis without a corresponding NEW process row; direct regulatory protein interactions support that biological synthesis, and a redundant process assertion was not added to satisfy a suggestion. All ten attached PMID quotations match the canonical records exactly, the seven PMID titles and availability flags match their caches, and the three recovered records are byte-identical to the verified Source28 archive and stage. The append-only session record is history/genes/human/ANKRD11/2026-09-27T214004Z-codex-d5af25.yaml.
The earlier final-adjudication statement that an MF coregulator annotation made a BP annotation redundant is superseded. GO:0003712 and GO:0006357 are different aspects. The current GO:0006357 definition and parents describe regulation of RNA polymerase II transcription: it is a child of GO:0006355 and regulates GO:0006366. Its positive and negative regulation children are not added alongside the unsigned parent. None of the fourteen original rows already represents this process or a transcription-regulation ancestor/descendant. The local GO-CAM index still has no ANKRD11/Q6UB99 match; this is a local coverage check, not a claim about all live models.
The participation test is satisfied by ANKRD11's own regulatory protein interactions and experimentally mapped repression/activation domains, considered together with endogenous perturbation. I reread the complete available main Methods, Results and Discussion of PMID:17521611(https://pmc.ncbi.nlm.nih.gov/articles/PMC1950474/). The Gal4 assays test tethered regulatory domains and an RD2 deletion; they do not establish native DNA-binding specificity. The separate endogenous shRNA experiment increases progesterone-receptor-dependent MMTV reporter output, with dose dependence and a second shRNA reported to give similar results. The separately checked original Figure 4C caption on the official PubMed article page specifies human HeLa cells transfected with hPR-B, the MMTV-luciferase reporter and ANCO-1 shRNA; COS-7 refers to the domain assays. The canonical extraction lacks these full captions, so the host assignment comes from that external primary caption, independently checked by the reviewer. Exact cached Results: “ANCO-1 shRNA was capable of enhancing PR-mediated transcriptional activation by 3 to 5-fold.” This is regulatory work at a transcriptional output, not inference from developmental necessity alone. The model of simultaneous p160/HDAC recruitment remains a model; no partner enzyme activity is attributed to ANKRD11. The complete PMID:18840648 abstract also reports that ANKRD11 silencing reduces p53 activation of CDKN1A. Its full paper was not read. One NEW GO:0006357 row uses IMP/PMID:17521611 and the unsigned term to summarize these partner-dependent directions.
The comparator check selected transcriptional coregulators in the same regulatory role. The current primary NCOR1 UniProt record, entry version 243 dated 2026-09-02, records MF GO:0003714 with IDA and BP GO:0000122 with IMP. The MGI/Alliance comparative table for NCOA1 explicitly displays human GOA experimental rows: MF GO:0003713 with IDA and BP GO:0045944 with IDA. That comparison graph is dated 2023-03-10, so it establishes the annotation convention historically rather than a fresh complete QuickGO census. HPA v21 SIN3A additionally displays GO:0003714 and GO:0000122 for human Q96ST3; its imported table does not expose the underlying evidence rows and is only historical corroboration. A single normal QuickGO batch request failed DNS; interactive AmiGO associations were unavailable. Failed access is not interpreted as absence. The two experimental comparator records demonstrate that a coregulator MF and transcription-regulation BP can coexist; ANKRD11's direct evidence, not comparator presence alone, justifies the addition.
For the retained PMID:25556659 nucleus row, the mouse chromatin/HDAC3 quotation was removed and replaced by the already verified human HPA location statement. The original human Figure 2 observation remains explicitly an external primary-source assessment from the initial review, separate from the abstract-only cache. The PMID:21782149 construct-species hedge remains: a donor laboratory's identity does not establish a construct sequence. No original source object or action changes.
The optional binding suggestions were considered without expanding this revision into a new interaction catalogue. The HDAC and p160 interactions inform the established coregulator mechanism; a dedicated HDAC-binding addition awaits full original interaction-assay assessment, and a separate p160-binding annotation is not required to describe the demonstrated regulatory activity. The direct ANKRD11–STAG2/RAD21 findings remain visible in the description and reference findings; choosing an appropriately specific interaction term remains open, rather than adding generic protein binding or inferring cohesin membership, ATPase activity or loop extrusion. The retained overlapping head/face developmental annotations describe related outcomes, not independent core molecular mechanisms; their original source-specific assertions and non-core decisions remain intact.
Prospective independent scientific consultation agreed that the direct regulatory evidence supports this single unsigned process addition. This follow-up preserves the published history and uses a new append-only session record. Validation and retrospective consultation are documented in the follow-up history.