The normal recovered seed contains 34 GO assertions and seven alternative products for human O75179. The original UniProt and GOA files, all source assertion fields, and alternative products are preserved. The baseline is recorded in tmp/ANKRD17-initial/baseline/manifest.json. The six original PMID caches were read. Two missing original caches (17276651 and 23711367) were imported by the parent from the verified seed7 artifact; the four pre-existing caches were preserved, including 19150984 and 22658674 whose staged copies differ.
The installed deep-research client was invoked through just deep-research-falcon human ANKRD17 --fallback perplexity-lite, with workspace-local UV directories and the absolute installed client override. It terminated with exit 1 in 5.346 seconds: Falcon authentication could not reach the network, and the fallback also failed with DNS errors. No provider report was produced or authored manually. The terminal receipt and log are in tmp/ANKRD17-initial/.
PMID:19150984, Identification and functional analysis of a novel cyclin e/cdk2 substrate ankrd17 supplies the main mechanistic evidence. The canonical cache contains the abstract, Results and Discussion; Methods were read separately in the verified staged full-paper extraction. These are distinct access scopes, and canonical supporting quotations must occur in the preserved canonical cache.
The historical human construct is the 2352-amino-acid isoform b, whereas the current UniProt canonical protein has 2603 residues. Residue numbering from that construct must not be silently assigned to the current canonical sequence or to one of the seven present isoform identifiers. Human 293T, U2OS and HeLa cells were used. The GST fragment was produced in bacteria and incubated with lysate containing transfected Myc-cyclin E/CDK2; this is not a purified two-component affinity measurement. Endogenous co-immunoprecipitation and mutation of the RQL cyclin-docking region also support association. ANKRD17 is a cyclin E/CDK2 substrate, not a demonstrated kinase or kinase activator.
Cell fractionation and immunostaining support a chromatin-associated nuclear pool. The insoluble fraction resists micrococcal nuclease but is salt released in a pattern comparable to replication proteins rather than the nuclear matrix control. These results support the existing chromatin and chromatin-binding assertions; they do not establish sequence-specific DNA recognition. Chromatin association changes during the cell cycle. Knockdown with independent siRNAs reduces DNA synthesis/S-phase entry, and overexpression increases S-phase entry, supporting the existing regulatory BP terms. Changes in p53/p21 abundance are not evidence that ANKRD17 directly binds their promoters.
The paper reports co-immunoprecipitation of MCM3, MCM5, MCM7, CDC6 and PCNA with tagged ANKRD17. Endogenous confirmation is reported for MCM7, CDC6 and PCNA. These assays show association; they do not identify every binary interface or establish ANKRD17 as an obligate replisome component. Knockdown reduces CDC6 abundance and chromatin-bound PCNA, while chromatin loading of the assayed MCM proteins is unchanged. The Discussion expressly leaves direct versus indirect PCNA-loading regulation unresolved. Accordingly, neither DNA clamp-loader activity nor replication-complex adaptor activity follows from these experiments alone.
PMID:22328336, Ankrd17 positively regulates RIG-I-like receptor (RLR)-mediated immune signaling was read as a complete abstract. The publisher full-text endpoint failed; no unobserved Methods, organism or construct detail is supplied from inference. The abstract explicitly reports association with RIG-I/DDX58, MDA5/IFIH1 and VISA/MAVS, enhanced association of the receptors with VISA, an ankyrin-domain requirement, and positive perturbation effects on RLR signaling. This is stronger evidence for an organizing signaling-adaptor activity than the separate replication co-immunoprecipitations. The NF-kappa-B and receptor-specific regulatory annotations describe this experimentally supported signaling role. A general RNA-binding role must not be confused with direct viral-RNA sensing by RIG-I/MDA5.
PMID:23711367, A role for the Ankyrin repeat containing protein Ankrd17 in Nod1- and Nod2-mediated inflammatory responses was read as a complete normal cached and official PubMed abstract. It reports N-terminal binding to NOD2, perturbation effects in human myeloid and epithelial cells, and inflammatory responses to Shigella in HeLa cells. It also reports that ANKRD17 did not affect Sendai-virus-induced type-I interferon in HeLa. That contextual result is retained alongside the RLR study; it is not a basis for blanket rejection of the RLR findings in their experimental context. NOD2 binding supports receptor binding without establishing ANKRD17 as the microbial ligand or naming a second bridged NOD-pathway partner. The full publisher body was inaccessible.
PMID:17276651, The VP1 structural protein of enterovirus 71 interacts with human ornithine decarboxylase and gene trap ankyrin repeat was read as a complete cached and official abstract. Human GTAR/KIAA0697 interaction was tested by yeast two-hybrid, co-immunoprecipitation and confocal microscopy in transfected and EV71-infected Vero cells. The abstract does not show the exact nuclear/cytoplasmic localization panels; Vero cells are not human cells. Independent human localization evidence can corroborate the broad terms without pretending those panels were read.
The official Human Protein Atlas ANKRD17 subcellular page, accessed 2026-09-27, states: "Mainly localized to the nucleoplasm. In addition localized to the nuclear membrane." Nucleoplasm is the supported main location, nuclear membrane an approved additional location, with antibody HPA063731 and human CACO-2, Rh30 and U2OS imaging. Its overall cell-cycle abundance result is not a measurement of the chromatin-bound fraction studied in 19150984. A new nuclear-membrane annotation would add overlapping coverage to existing nuclear location assertions and is not proposed.
PMID:19946888, Defining the membrane proteome of NK cells was read as a complete abstract. The YTS membrane preparation contains many proteins without predicted membrane-spanning structure, including transient membrane-associated proteins. The ANKRD17 peptide/supplement row was not recovered. The membrane HDA assertion therefore remains unresolved; nuclear localization does not refute an additional peripheral membrane pool.
PMID:22658674, Insights into RNA biology from an atlas of mammalian mRNA-binding proteins was read as a complete abstract. The experiment uses UV crosslinking and mRNA-interactome capture in human HeLa cells, but the exact ANKRD17 supplementary row was not independently recovered. The KH domain and the curator's HDA assertion are compatible with RNA binding. They do not establish an RNA sequence preference, an RNA-processing mechanism or catalytic activity. Access limitations are recorded explicitly for this experimental source.
At the initial research stage, five additional PMID identities and substantive primary-source content were checked while normal cache closure was pending. The Source29 section below records the subsequent cache closure and expanded read scopes.
Official AmiGO definitions were checked for signaling adaptor activity GO:0035591, signaling receptor binding GO:0005102, protein kinase binding GO:0019901, ATPase binding GO:0051117, and RNA polymerase II transcription regulator complex GO:0090575. The last denotes a complex acting at a regulatory DNA region, not merely a protein that affects YAP abundance or entry into the nucleus. The local GO index contains no generic helicase-binding or PCNA-binding term suitable for an automatic replacement of all replication co-immunoprecipitations. GO:0017151 is specifically DEAD/H-box RNA-helicase binding and is inappropriate for MCM DNA helicases. GO:0003689 DNA clamp-loader activity requires ATP-driven opening/closing of the clamp around DNA and is not supported here.
The two IBAs share PTN001506277, but this does not turn them into equivalent biological claims. ANKRD17's appearance among the innate-immunity descendant evidence is legitimate and is not circular. The Pol-II complex assertion includes fly Mask evidence. A fly donor is not by itself evidence of a propagation error. The family tree, IBD node placement and exact complex evidence need to be examined before assigning a confident phylogenetic failure; their current unavailability is an uncertainty, not a demonstrated lineage-specific loss.
No NEW biological-process assertion is planned. Existing regulatory processes cover the direct work supported by the experiments; disease, knockout necessity, altered downstream expression and unverified complex membership are not grounds for additional process annotations.
The preserved UniProt record explicitly identifies IPR004087 (KH domain), IPR004088 (type-1 KH domain), IPR036612 (type-1 KH superfamily), and PROSITE PS50084. Direct InterPro pages/API were inaccessible. The official PROSITE PDOC50084 documentation was read and identifies KH domains as RNA-binding modules. Its cited bibliography was not separately consulted. This supports the domain-based RNA-binding inference without claiming that the ANKRD17-specific HDA supplemental record was read.
One normal fetch-panther-paint PTHR23206 --extra-uniprot O75179 attempt used the verified recovered member records copied into an isolated tmp output directory. It exited 1 in 3.909 seconds on DNS failure for data.pantherdb.org; no PAINT slice was produced and no shared cache was written. The exact node, family tree and MSA remain unavailable. The five additional publication records received one ordinary fetch-pmid batch attempt in an isolated output directory; it reached its 45-second bound with zero outputs. This is a batch timeout, not five independently completed fetch failures. Exact identity and terminal receipts are in tmp/ANKRD17-initial/ and the records are queued for Source29 recovery.
The initial authored draft reviews all 34 assertions: 20 ACCEPT, 7 MODIFY, 4 KEEP_AS_NON_CORE and 3 UNDECIDED, with two core functions and no NEW assertions. All seven alternative products and all source assertion objects are preserved. Independent preliminary consultation prompted retention of the MCM3/5/7 co-immunoprecipitations as non-core generic associations: shared-complex recovery does not resolve an ATPase-specific interface. CDC6 is separately refined as binding to the named replication ATPase, without claiming a direct purified interface or stimulation. The direct substrate/kinase evidence supports the CDK2 refinement. Full authored consultation and five-source cache closure remain pending.
Initial full gene validation and rendering passed. Five advisories were recorded at that stage; the protein-binding interpretation below is superseded by the PR3340 followup: four bare protein-binding associations remain non-core because the actual experiments support association but not a defensible informative replacement; removing a valid experimental assertion solely for low information would not follow the project action definition. The RNA-binding action difference is source-specific: type-1 KH-domain inference is supported, while the experimental HDA supplement remains unread. These are not evidence that the experimental associations are false. A new history record was scaffolded with the normal helper; no published history was changed.
All five additional normal records were independently identified and read within the scopes below, then imported exclusively by the coordinator. Canonical bytes match the verified run 36354596463 artifact and staging records; none of the six earlier caches or raw seed records was rewritten. All eleven consulted PMIDs are now in the bibliography and normal cache. A full-text metadata flag does not mean every paragraph or supplement was inspected.
All 34 source assertions and seven alternative products remain intact. Decisions remained 20 ACCEPT, 7 MODIFY, 4 KEEP_AS_NON_CORE and 3 UNDECIDED, with zero NEW and two core functions. The three uncertainties are the unread target-specific RNA-interactome supplement, the unread NK-cell membrane-proteome supplement, and the unresolved PAINT polymerase-II-complex ancestral node. These are biological/source-resolution limits, not missing normal cache files. Completion records validation, rendering, independent consultation and exact source-object/citation checks separately.
The published baseline at 03ce48fc6a707db7080da9614b43842c754722e9 was copied and verified before this revision. All 34 source assertions, seven alternative products, raw inputs, eleven PMID caches and published history remain unchanged. The earlier four generic-binding non-core choices are superseded here.
The actual PMID:19150984(https://pmc.ncbi.nlm.nih.gov/articles/PMC2658080/) Results for figure 7D report FLAG-ANKRD17 co-immunoprecipitation of MCM3, MCM5, MCM7, CDC6 and PCNA in human 293T cells; figure 7E additionally confirms endogenous MCM7, CDC6 and PCNA associations. These are positive association results, with direct versus indirect interfaces unresolved. The revision consistently refines all four ATPase-partner rows (MCM3/5/7 and CDC6) to ATPase binding, GO:0051117, which is an enzyme-binding term. It describes partner class at the existing IPI evidence level, not an ATPase active-site contact, stimulation, intrinsic ANKRD17 ATPase activity or a loading reaction.
The current reviewed human MCM7 record P33993 identifies MCM3 and MCM5 in the MCM2-7 ring and places the six ATPase sites at neighboring-subunit interfaces. This provides collective motor-subunit class context; it does not imply autonomous ATPase activity of an isolated MCM subunit. Individual MCM3/P25205 and MCM5/P33992 live entry and REST pages were inaccessible in this followup, so no individual-record reading is claimed. NCBI CDC6 identifies the AAA ATPase class, while human PCNA P12004 is a DNA sliding clamp. The PCNA generic-binding row is removed as uninformative under the touched-annotation policy, without disputing its positive co-immunoprecipitation. A clamp-loader or replication-adaptor replacement remains unsupported; the paper expressly leaves the PCNA-loading mechanism unresolved.
The actual cached PMID:31661072(https://elifesciences.org/articles/48601) figure-1 supplement caption states: "Mask has no intrinsic transcriptional co-activator activity in a GAL4-UAS reporter assay." This is Drosophila S2 GAL4-DBD-Mask versus DBD/Yki reporter evidence. Separate fly NLS/rescue experiments support Yki import. Neither finding alone settles the GO:0090575 complex assertion: lack of intrinsic activation does not exclude physical membership in a regulatory-DNA complex. The IBA remains UNDECIDED because the exact PAINT node and donor complex evidence are untraced, with no donor-count or species-based rejection. No new phylogenetic access is claimed.
Two reference findings now attach existing biological-summary claims to their sources. The figure-6 Results of PMID:31661072 specifically report increased YAP protein levels after Mask2 overexpression in human Caco2 cells; the separate large-droplet HEK293T assay used Mask1. The complete cached abstract of PMID:33909992(https://pubmed.ncbi.nlm.nih.gov/33909992/) describes 34 individuals from 32 families and a variant spectrum supporting haploinsufficiency. Its relevance is HIGH for the human disease summary, while the abstract-only limit and absence of a defined developmental molecular step remain explicit. This followup adds no GO assertion or core function.
The unread target-specific NK-cell membrane-proteome evidence remains UNDECIDED: absence of a transmembrane segment does not refute peripheral association. Revised decisions are 20 ACCEPT, 10 MODIFY, one REMOVE and three UNDECIDED, with two unchanged core functions and zero NEW assertions. Historical read scopes above are retained and clarified; current quotation, source-integrity, schema, history and rendering checks are recorded separately.