Human ALPK1 is HGNC:20917 / UniProtKB:Q96QP1 (aliases LAK/Lak, KIAA1527 and FLJ22670). The normal seed5 import supplies 47 PENDING GO assertions, 24 original references and two alternative products. The exact three-file baseline is recorded in tmp/ALPK1-full-review/baseline.json. All annotation source objects, qualifiers, supporting entities and products are preserved. Seed5 current-main canonical/alias and open-PR preflights were checked by the coordinator; a fresh publication preflight remains necessary because main advances during this campaign.
The default deep-research launcher and the supported installed-client retry failed; the exact terminal logs and matching client version are recorded in tmp/seed5-research-attempts-receipt.json. No provider report was authored or fabricated. This file is the manual research record. The six original cited PMID records and twelve Reactome records are present. Three original publications have full text in the normal cache (27169898, 28222186, 30967659); 28877472, 30111836 and 36931259 remain abstract-only. External reading is identified separately below and does not change the immutable cache flags.
ALPK1 couples microbial nucleotide-heptose recognition to TIFA phosphorylation. Its N-terminal domain recognizes ADP-heptose and activates the C-terminal alpha-kinase; phosphorylated TIFA assembles signaling oligomers that engage TRAF6 and downstream TAK1/IKK/NF-κB machinery. This is direct sensing and catalytic participation, not a role inferred merely from knockout necessity. Human cellular/biochemical results and mouse inflammatory phenotypes provide different levels of evidence [PMID:30111836; PMID:28222186; PMID:28877472].
One integrated kinase/sensing core captures the strongest mechanism. Cytosolic localization is core; ciliary-base, centrosomal and mitotic spindle-pole pools are real but their local biochemical substrates remain uncertain. The broad innate-response assertions remain ACCEPT because they describe that same core biology. They are not demoted solely because a more specific pathway term also exists. No NEW annotation row is proposed. No new process is introduced into the core: both processes were already seeded. Generic binding is not used as a stand-alone function.
The cached complete Results/Methods and Figure 4 were inspected, with an independent bounded peer read. Human ALPK1 constructs were expressed in HEK293F/HEK293T systems. FLAG-myosin IIA immunoprecipitated from HEK293T cells was phosphatase-treated and then incubated with ATP and recombinant ALPK1. ProQ staining and omission controls support phosphorylation. ALPK1 purification also coeluted endogenous myosin, explicitly acknowledged in the paper, so this is not described as a two-purified-component reconstitution. Detection by ProQ and pan-phospho-Ser/Thr does not identify a serine acceptor site. The source names MYH9/nonmuscle myosin IIA; the immutable UniProt function paragraph's MYO9A wording is not propagated.
THP1 cells treated with monosodium urate and human gout/control expression comparisons provide contextual evidence for inflammatory secretion. Golgi-derived transport is an interpretation rather than a directly tracked ALPK1 vesicle-transport reaction. The positive kinase assay is retained; the seeded serine-only annotation remains UNDECIDED instead of implying ALPK1 cannot phosphorylate serine.
The linked correction PMID:27283228, DOI 10.1038/srep27323, was verified through official PubMed and the indexed original PMC4901552 notice. It corrects Chia-Lin Chen's affiliation only, to the Genomics Research Center, Academia Sinica. It does not change experiments, results or conclusions. The normal cache fetch failed; this remains a required citation gate.
The complete cached 28222186 Results, Figure 6 and Methods establish a human-cell pathway: ALPK1 depletion reduces infection-induced cytokines, IκBα degradation and p65 translocation; siRNA-resistant full-length ALPK1 restores TIFA oligomerization while a kinase-domain deletion does not. The rescue and upstream placement are positive evidence of signaling work. HeLa and Caco-2 systems are distinguished from the infecting bacterial species. ALPK1 was not required for the tested TNFα or PMA response, so the result is not generalized to every route into NF-κB.
The 28877472 local cache and independently verified PubMed abstract report functional ALPK1-dependent TIFA phosphorylation, TIFAsome formation and NF-κB activation during H. pylori infection. Full original assay-level methods were not independently recovered. Acceptance of its pathway annotations is supported by the positive abstract and convergent direct mechanism, rather than invented details of its constructs.
Both 2017 studies used an HBP-centered model. The 2018 work distinguishes direct ALPK1 recognition of ADP-heptose from host conversion of HBP into an activating nucleotide derivative. This refines the ligand mechanism without discarding the actual earlier signaling measurements.
The local record is abstract-only. External access consisted of the original first-page preview (page 122) through the author-uploaded article preview, and the publisher's abstract and Extended Data captions. It did not expose the complete original Methods or all main-text pages. The original preview carries © 2018 Springer Nature Limited, all rights reserved; no full paper is copied into the repository and no special full-text field is needed for this review.
The inspected material reports wild-type versus kinase-inactive K1067M rescue, TIFA Thr9 phosphorylation, ligand-dependent co-immunoprecipitation, and persistence of ALPK1's cytoplasmic localization. Extended Data describe N-terminal and kinase-domain coexpression, ADP-heptose identification by chromatography/mass spectrometry and binding by biolayer interferometry. The primary RCSB structure 5Z2C independently identifies human ALPK1 N-terminal protein expressed in E. coli; the bacterial host does not make the construct bacterial ALPK1. These data support direct receptor binding and kinase activity. They do not resolve the separate serine-only assertion, and incomplete full-text access is not evidence that the original curators were wrong.
The full cached Results, Methods and Figure 4 were read; the source identity was checked against PubMed. Endogenous ALPK1 staining in human ARPE19 cells localizes to mitotic spindle poles (Figure 4a–f), interphase centrosomes with diffuse cytosolic signal (4g–i), and the primary-cilium base after 24-hour serum starvation (4j–l). Mouse retinal connecting-cilium/basal-body staining in Figure 3 is a separate experiment and is not assigned human tissue identity.
Fibroblasts from two affected p.Thr237Met patients showed lower ciliation than one unaffected control, with three independent experiments scoring 100 cells per condition. HeLa overexpression of the variant increased multinucleation. These support the original contextual ciliogenesis IMP assertion and localization findings. They do not establish a dedicated cilium-assembly enzyme, a mapped centrosomal substrate or a loss-of-function mechanism. Later positive gain-of-function studies are considered below.
The Human Protein Atlas subcellular source independently supports centrosome and ciliary basal-body locations with HPA027435: centrosomal staining in A-431, U-251MG and ASC52telo, and basal-body staining in ASC52telo. Some tested lines lacked staining, including U2OS and serum-starved hTERT-RPE1; this is not interpreted as universal absence or used to override the ARPE19 experiments.
The official PubMed record verifies the title, DOI and human 14-3-3 interactome scope. The ALPK1-specific supplemental assay for the P62258 interaction was not recovered. This IPI row remains UNDECIDED; it does not make ALPK1 a chaperone or establish a localization-regulatory mechanism. By contrast, the accessible ALPK1–TIFA interaction is understood in its already annotated kinase mechanism; that same-source generic protein-binding row is removed as uninformative without denying binding.
The GO:0048029 monosaccharide-binding definition and ChEBI:59967 were inspected. ChEBI identifies the actual ligand as a nucleotide-sugar oxoanion. Recognition of an intact ADP-heptose molecule is not automatically evidence for binding either isolated ADP or free heptose. The two seeded monosaccharide-binding assertions are therefore MODIFY to GO:0097367 carbohydrate derivative binding, whose definition explicitly covers a carbohydrate derivative. This expresses the positive binding chemistry without claiming that free-heptose binding is impossible. A bounded independent peer found no convention requiring the narrower free-monosaccharide interpretation.
The three seeded serine-kinase assertions remain UNDECIDED because exact residue-class support is unresolved. Broad GO:0004674 kinase assertions remain ACCEPT. The live AmiGO record lists both reaction classes in its definition but also includes threonine-specific kinase synonyms, alternate ID GO:0106311, and threonine-directed histone-kinase children. These positive ontology features support retaining the conventional broad classification for the demonstrated TIFA Thr9 chemistry; they do not establish an unmeasured serine substrate or make GO:0106310 a demonstrated child of this term. RHEA:17989 is a valid serine-phosphorylation reaction, but its cited source resolution is a separate issue.
PAINT assertions are traced to seeded PANTHER:PTN002928819 only in structured propagation assessments. The extant human target appearing in source entities is legitimate descendant experimental evidence. Local PTHR46747 family metadata/entries exist, but a full PAINT tree/IBD file does not; no fabricated node-placement reconstruction is claimed. The binding refinement concerns chemical scope, not donor count, target self-evidence or lineage divergence. No ALPK1/Q96QP1 activity was found in the local GO-CAM index. The primary receptor/catalytic evidence already supplies the performer role for the existing core processes.
All twelve cached human Reactome summaries were read: ligand binding (9645428), TIFA binding/phosphorylation/oligomerization (9645524/9645535/9645481), TRAF6 recruitment/oligomerization/autoubiquitination/free-chain synthesis (9645520/9645501/9645414/9645394), TAK1 recruitment/autophosphorylation (9645406/9645442), IKBKG ubiquitination (9758604), and the parent pathway (9645460). The location assertions remain ACCEPT. None transfers TRAF6 ubiquitin chemistry or TAK1/IKK catalysis to ALPK1. These records are normal imported source caches; their internal bibliographies are not recursively treated as new authored citations when no independent claim relies on those papers.
All 47 source assertions are reviewed: 30 ACCEPT, 10 KEEP_AS_NON_CORE, 2 MODIFY, 1 REMOVE and 4 UNDECIDED; zero NEW rows. One integrated core and four biological questions are authored. All 24 original reference identities/titles and both alternative products remain exact. Five additional reference entries comprise four relevant publications plus the immutable UniProt file.
The notes-inclusive PMID census is ten distinct sources. Six original records are cached; four required records remain absent: 27283228, 37072480, 38060563 and 39116220. The coordinator's ordinary correction fetch failed DNS with cached 0/3 in tmp/seed5-additional-source-fetch.log; the three additional sources failed DNS with cached 0/3 and terminal exit 1 in tmp/ALPK1-new-source-fetch.log. All twelve cited Reactome caches are present. No provider artifact exists and no unrelated UniProt bibliography or unasserted search lead is added to this finite census. DRAFT reflects these real source-cache gates; identity verification and external primary reading are recorded independently from machine-cache fulfillment.
Full validation passed with one warning category for the four missing PMID caches. Integrity checks preserve all 47 source objects, 24 original reference identities and two products; all 75 cached quotations match case-sensitively after whitespace normalization, with no case-fold-only matches. History validation and rendering passed. Independent bounded source/core consultation found no biological blocker. Publication remains pending recovery and the coordinator’s final complete review/current-main check.
The four initial cache gates above are now closed by the exact normal-fetch source23 records. Source bytes were staged with strict archive/control checks and independently read before the coordinator imported them. Two records contain full scientific bodies (PMID:37072480 and PMID:38060563); PMID:39116220 contains the scientific abstract; PMID:27283228 is bibliographic-only with no notice body. The latter’s externally verified affiliation-only correction remains distinct from its local access flag. No original source assertion or action changed.
The recovered human kinase assay PMID:37072480 now directly supports the four retained broad kinase rows, the existing serine-specific uncertainty explanation and the integrated core through its exact Thr9 result. Human myc-ALPK1 was immunoprecipitated from HEK293; GST-TIFA was produced in E. coli. ATP/pThr9 and ATPγS assays include kinase-dead K1067R, domain-deletion and reagent/enzyme omission controls. These are strong catalytic data with the preparation limits of an immunoprecipitate. Additional threonine-site results and kinase-dependent ALPK1 phosphorylation do not manufacture a mapped serine site or settle cis/trans autophosphorylation.
The recovered PMID:38060563 body preserves mutant-specific activation by selected human nucleotide sugars and explicitly leaves wild-type binding without activation unresolved. Its T237M activity relative to wild type differs from the differently configured thiophosphorylation assay in PMID:37072480; retain each assay’s conditions rather than assert a universal mutant rate. Y254C purification lability limits its in-vitro ligand assignment. Neither study directly identifies the nucleotide sugar occupying mutant ALPK1 in patients. PMID:39116220 remains scoped to its abstract and the separately inspected deposited human structure; full Science Methods were not recovered.
The final finite citation census remains ten PMIDs and twelve Reactome records, all present; there is no provider report and no fabricated research output. The initial reviewed snapshot, source23 scientific assessment and normal import receipt retain the recovery history. All 47 source objects, 24 original reference identities, two products, actions and core term assignments remain unchanged; six exact cached Thr9 evidence attachments were added. The newly authored correction title now matches the normal record’s terminal period.
Final full validation completed successfully with no curation warnings. History validation and rendering passed. The integrity audit preserved all seeded source fields/products and original reference identities; all 81 cached quotations match case-sensitively after whitespace normalization, with zero case-fold-only matches. The rendered page contains the new evidence and no duplicated PubMed-link wrappers. Status is COMPLETE; publication remains subject to the coordinator’s final review and fresh-main path check.
This follow-up addresses formal review 5331472772 and full comment 5858370723 against published head fb6a4aad9c660bf2b4974f84322442edceb85f60. The judgments below supersede the earlier ligand-replacement and ciliogenesis conclusions while preserving the positive source findings.
The live official GO:0038187 definition and hierarchy describe recognition of a microbial pattern that initiates an innate immune response. Its parent is signaling receptor activity, without a transmembrane restriction. The live ATP-binding hierarchy confirms GO:0097367 carbohydrate derivative binding as an ancestor. That broad chemical classification is true but does not add the receptor mechanism beyond ALPK1's retained nucleotide binding. Both existing monosaccharide-binding rows now propose pattern recognition receptor activity. This is a source-supported functional refinement of existing assertions, not a new annotation or a claim that PAINT placed the function at the wrong ancestral node. No ADP-heptose-specific child was established in the inspected hierarchy, and no new term is invented. The local GO-CAM index contains no Q96QP1 entry; that absence is not treated as a missing annotation.
For PMID:30111836, the official PubMed abstract and original Nature Extended Data captions were reread. Direct ADP-heptose binding, ligand-stimulated TIFA phosphorylation and wild-type versus kinase-inactive human-cell rescue positively establish receptor activity. Extended Data distinguish human ALPK1 constructs from bacterial expression hosts; RCSB 5Z2C independently identifies human ALPK1 N-terminal protein expressed in E. coli BL21(DE3). The captions also report HBP binding without sufficient activation, so ligand binding alone is not equated with agonism. The local cache remains abstract-only, and the publisher preview/captions are not described as recovered full main Methods. The existing exact cached abstract quote remains the machine-readable support.
The complete cached PMID:30967659 Results, Figure 4 and Methods were reread. The ciliation experiment compares two T237M patient fibroblast lines with one unaffected control after serum starvation, across three experiments scoring 100 cells per condition. Endogenous ARPE19 localization at the cilium base is positive location evidence. The HeLa wild-type/mutant comparison instead measures multinucleation. Neither experiment resolves normal ALPK1's participation in cilium assembly against a secondary consequence of variant signaling; no wild-type loss/rescue ciliation assay was identified in this source. The paper itself raises a gain-of-function interpretation, and the separate mutant kinase studies PMID:37072480 and PMID:38060563 establish signaling changes without resolving normal ciliogenesis. The existing cilium-assembly IMP row is therefore UNDECIDED. This preserves the real patient defect and curator assertion, does not reject upstream signaling as a possible means of participation, and does not require a dedicated ciliary enzyme. The biological question now asks how to distinguish these mechanisms.
The two identical duplicated evidence entries were removed from the same-source NF-kappaB and innate-response rows. The optional serine-kinase suggestion does not resolve a mapped serine acceptor: PMID:27169898 already supports a retained broader kinase assertion, while its exact serine-specific assertion remains UNDECIDED. Qualifiers are unchanged. No mechanical downgrade of the twelve broad but valid cytosolic Reactome assertions was made.
The resulting tally is 30 ACCEPT, 9 KEEP_AS_NON_CORE, 2 MODIFY, 1 REMOVE and 5 UNDECIDED across the same 47 source assertions; zero NEW rows. All 29 reference identities, two products, the integrated core and source bytes are preserved. The finite citation census remains ten PMIDs and twelve Reactome records, all cached, with no new recovery request or fabricated provider report.
Follow-up verification: full ordinary validation passed with no curation warnings; history and rendering passed. All 79 remaining cached quotations match case-sensitively after whitespace normalization, with no case-fold-only matches. Parsed checks preserve all source fields, reference objects, products and the integrated core; raw caches and published history hashes are unchanged. The rendered evidence is present without duplicate PubMed-link wrappers. Status remains COMPLETE. The coordinator retains publication and independent review.