Human ALK is HGNC:427, UniProt Q9UM73, a 1,620-residue receptor. The approved-symbol snapshot includes CD246 and ALK1; ALK1 is also an ambiguous historical name and must not be used to conflate this protein with ACVRL1. Canonical and alias preflight found no overlapping gene review. The normal seed5 fetch produced the immutable UniProt/GOA records and 110 seeded annotation objects, all initially PENDING. The two ordinary autophosphorylation relationship variants are retained as distinct source objects. No source identifiers, terms, evidence codes, relations, supporting entities or isoform fields were rewritten. No alternative-product objects were supplied by this seed.
The seed5 canonical and auxiliary import receipts establish the exact normal-fetch origin of the source files. Existing cache variants were preserved during auxiliary import; no merged or authored publication cache was manufactured. This review used the available canonical version, with actual body access checked independently of its metadata flag.
The first normal deep-research launcher failed on the tool-cache permission boundary. Root then invoked the same installed deep-research client (0.2.7rc1) through the normal wrapper with DEEP_RESEARCH_CLIENT_CMD=.venv/bin/deep-research-client. Falcon and the configured perplexity-lite fallback both failed on DNS before yielding a report. Their actual logs are tmp/ALK-seed5-deep-research.log and tmp/ALK-seed5-deep-research-installed.log. There is no genuine provider report to cite. These are manual research notes, not a provider artifact.
All seeded publication and Reactome records were read. Manual research included primary PubMed records, indexed primary Methods/Results, two original-paper PDF routes, live GO definitions, the Human Protein Atlas subcellular page, the MGI comparative Alk graph, and the cached PAINT IBD table. Failed normal cache retrieval and successful external primary-text access are different events. External access does not set a missing local cache to available.
The central activity is a transmembrane receptor tyrosine kinase. Normal human ALK binds extracellular ligands, assembles receptor complexes, and performs intracellular ATP-dependent tyrosine phosphorylation. Its broad kinase mappings are refined where the source and human receptor evidence support that subtype. Reaction models built partly from NPM–ALK or isolated kinase activity retain tyrosine kinase chemistry at their measured resolution; they do not imply that a fusion lacking the normal extracellular/transmembrane region is itself a membrane-spanning receptor.
PMID:9174053 identifies a full-length normal human cDNA and surface-labeled glycoprotein, while its developmental expression survey is mouse. The cached abstract explicitly states: “Surface labeling studies indicated that the 200 kDa glycoprotein is exposed at the cell membrane”. That distinction supports location without turning a mouse expression map into a human developmental perturbation assay. Full original Results were not recovered.
PMID:26418745 directly tests human ALK recognition of FAM150A/B (ALKAL1/2), including purified receptor ectodomain-Fc binding experiments, cellular phosphorylation and differentiation. Human proteins in rat PC12 or fly hosts are distinguished from endogenous human neuroblastoma experiments. The Fc tag in a purified binding reagent is not an oncogenic ALK fusion. PMID:30061385 compares full human ALKAL2/augmentor-alpha with a biologically active ligand fragment. Its ligand disulfide dimerization experiments do not automatically demonstrate receptor self-association; receptor phosphorylation and differentiation in human NB1 versus engineered rodent systems are separate measurements.
PMID:34646012 and PMID:34819673 provide actual human receptor–cytokine structures and receptor interface experiments. The former tests a receptor–receptor site 3 mutation, M751T, in Ba/F3 cells with expression controls. The latter examines extracellular dimerization mutants and full-length receptor phosphorylation in NIH3T3 cells. These are positive receptor–receptor results, not ligand-only papers. However, soluble construct/ligand stoichiometries differ between experiments, so no universal native 2:1 or 2:2 complex is asserted. ALKAL2 is a potent ALK activator; ALKAL1 responses depend on preparation, concentration and assay context. Equal physiological potency is not inferred from activation in an engineered system.
PMID:25605972 directly supports extracellular heparin binding at abstract-access resolution: short chains bind without activating, whereas longer chains promote receptor dimerization/activation in cultured neuroblastoma cells. Heparin is integrated into the receptor mechanism; it is not treated as proof that every glycosaminoglycan has the same action. PMID:17681947 explicitly reports that PTN can activate ALK without direct PTN–ALK contact, through inhibition of the PTPRZ phosphatase. This positive indirect mechanism is distinguished from older direct PTN/MDK ligand models. It does not establish that all possible direct interactions are impossible.
PMID:25517749 provides human neuroblastoma mutation data, purified human kinase activity/ATP kinetics and full-length cellular transformation measurements. Its main text did not expose the exact self-binding experiment used by the seeded IPI row. Independent receptor–receptor structural evidence supports the self-association assertion, with that original-source gap stated explicitly. PMID:19459784 is a full review, useful for separating normal receptor, developmental models, oncogenic fusions and proposed dependence-receptor behavior. It predates discovery of ALKAL ligands. Caspase cleavage in a proposed dependence-receptor model does not make ALK a protease, and cellular survival effects do not imply a uniform apoptotic direction in every tissue.
The live definitions of GO:0004714, GO:0007169, GO:0038083 and GO:0043235 support an integrated receptor kinase/signaling core. GO:0038083 includes autophosphorylation between identical receptor polypeptides; “auto” does not restrict it to cis phosphorylation. ATP binding, self-association and heparin binding are valid parts of this integrated mechanism, not separate unsupported core functions.
GO:0030298 receptor signaling protein tyrosine kinase activator activity inherits the noncovalent modulation requirement through GO:0140677 and GO:0098772. The positive receptor interface data are retained, but the observed phosphorylation/proliferation responses do not isolate a separable noncovalent increase in kinase activity from trans-autophosphorylation. The two experimental rows and ARBA row therefore remain UNDECIDED; the reason is mechanistic/ontology scope, not “ALK is a receptor, therefore it cannot activate another receptor”. No separate activator core is asserted.
Broad membrane and plasma-membrane assertions remain valid at their source resolution. The current Human Protein Atlas subcellular page reports supported plasma-membrane staining with HPA010694 in Rh30 cells. The displayed SH-SY5Y experiment has no staining; positive Rh30 evidence is not generalized to every cell line.
Generic interaction rows are assessed source by source. The functional human ALKAL interaction in PMID:26418745 is refined to the established receptor activity. The NCK2 co-immunoprecipitation result is explicitly positive in the abstract of PMID:34606829 and retained NON_CORE at generic binding resolution. Exact ALK–HSP90AB1 PMID:22939624, SHC1 BioPlex PMID:33961781, and the PIK3R1/SHC1/RET screen pairs [PMID:34606829; PMID:35384245] remain UNDECIDED where supporting-data experiments were not recovered. A full-text flag is not evidence that the particular pair-level assay was read. No generic interaction was declared false merely because a more informative function would be desirable.
The precise complex intended by the PX-FERM trafficking annotation PMID:23382219 and ALK target/peptide detection in urinary exosomes PMID:19056867 remain unresolved. A paper focused on other cargos can contain a valid ALK experiment; neither an absent title mention nor cell-surface receptor biology establishes misattribution or contamination.
No NEW annotations are proposed. The cached GO-CAM index has no Q9UM73 or mouse P97793 entry. The review does not infer omissions from that absence, and does not manufacture process participation from disease association or downstream necessity. The core contains the existing receptor activity and direct signaling/autophosphorylation steps; contextual developmental, apoptotic and systemic metabolic roles are retained separately.
The PTHR24416 PAINT table was inspected, including matching IBD entries at PTN001230349 for receptor kinase, plasma membrane, signaling and receptor complex. The target's appearance among descendant evidence is legitimate experimental grounding, not circularity. The developmental nodes PTN000699407 and PTN000699408 explicitly trace to zebrafish and mouse evidence, respectively. A short donor list is not treated as weak evidence. The full tree/MSA was not independently reconstructed, and no target-specific loss or misplaced clade is claimed.
The official MGI comparative Alk GO graph resolves the following term-specific donor sources. Model and construct scope were checked rather than inferred from the host species.
| Primary source | Actual evidence read and interpretation |
|---|---|
| PMID:32442405 | Identification of ALK in Thinness, DOI 10.1016/j.cell.2020.04.034. Primary PubMed identity and the original Cell PDF hosted by an author were read, including Results/Figures 5–6 and Methods. Global and PVN-specific conditional mouse Alk loss increase energy expenditure and adipose sympathetic/lipolytic output. Adult conditional deletion and peripheral tissue controls support a neural regulatory mechanism. ALK is not the lipase. Human association is not a human perturbation assay; the reported thinness indels did not replicate in the UK Biobank analysis. |
| PMID:30497772 | Neural Stem Cells Behave as a Functional Niche for the Maturation of Newborn Neurons through the Secretion of PTN, DOI 10.1016/j.neuron.2018.10.051. PubMed identity/abstract and indexed original Results/Figures 5–6 from the author-institution PDF were read; direct whole-PDF retrieval timed out. Mouse newborn dentate-neuron Alk knockdown reduces dendrite length/complexity and alters PTN/AKT responses. Functional PTN dependence is not a direct physical ligand-binding assay. |
| PMID:11121404 | Activation of anaplastic lymphoma kinase receptor tyrosine kinase induces neuronal differentiation through the mitogen-activated protein kinase pathway, DOI 10.1074/jbc.M007333200. Primary PubMed abstract describes a mouse-IgG2b-Fc/ALK chimera in PC12 cells, kinase-inactive/wild-type controls and MAPK-dependent differentiation. The complete original Methods were not recovered. Although the graph links rat developmental evidence, rat PC12 host identity does not prove the ALK construct's species. |
| PMID:17487225 | Behavioral and neurochemical alterations in mice deficient in anaplastic lymphoma kinase suggest therapeutic potential for psychiatric indications, DOI 10.1038/sj.npp.1301446. Primary PubMed abstract reports increased basal hippocampal progenitor proliferation after mouse kinase-domain deletion, with neurochemical/behavioral phenotypes. It supports unsigned regulation, not a universal positive proliferation role. Full body was not recovered. |
| PMID:23667670 | Anaplastic lymphoma kinase is required for neurogenesis in the developing central nervous system of zebrafish, DOI 10.1371/journal.pone.0063757. Primary PubMed/PMC3648509 and indexed original Methods/Results describe zebrafish morpholino/inhibitor and heat-shock overexpression experiments. Progenitor formation, neuronal differentiation, survival and positioning were distinguished. This is donor-species functional evidence, not a human construct experiment. |
| PMID:41771741 | Ribosomal Profiling of the Geniculate Ganglion Identifies the Receptor ALK as Critical for the Development and Maintenance of Oral Sensory Neurons, DOI 10.1523/JNEUROSCI.1186-25.2026. Primary PubMed/PMC13055147 and indexed original Methods/Results were read. Mouse developmental knockout reduces taste-bud and innervation measures without geniculate-neuron loss; separate adult ceritinib experiments produce atrophy/death. These outcomes are not conflated, and no new human sensory-process assertion is proposed. |
These six papers are required by the explicit donor/source review. Normal fetch-pmid attempts were made once for the settled five donor IDs and once for the newly identified 2026 source. Both completed with exit 1, cached 0/5 and 0/1, respectively, due to DNS. No local cache was created. Durable receipts are tmp/ALK-full-audit/donor-normal-fetch-receipt.json and tmp/ALK-full-audit/current-normal-fetch-receipt.json, each recording its command, result and raw-log SHA256. The six source records remain required cache gates. External full-source access for some of them does not remove that repository gate.
All 42 distinct seeded Reactome records, including the four R-NUL mixed-species records, were read at their actual summary resolution. Their original IDs and titles are preserved. Plasma-membrane annotations are judged as location assertions, using the normal-receptor model and independent surface evidence, not by treating each event title as an ALK activity. Specific limitations are recorded in each reference assessment and relevant reason:
Internal bibliographies of cached pathway summaries and primary papers were not recursively claimed as read primary evidence. The citation census covers actual authored uses, direct donor tracing, notes and any genuine provider artifacts; it excludes untouched raw bibliography-only UniProt records that this review does not rely on.
The original PMID:9174053 abstract and UniProt both identify an erratum at Oncogene 1997 Dec 4;15(23):2883. Targeted primary PubMed, title/bibliography and Nature issue searches did not recover its content or a verified PMID/DOI. The erratum is therefore an explicit unresolved source-access question. No identifier was guessed, no synthetic correction cache was created, and no assertion was made that the correction is harmless, invalidates the paper, or constitutes retraction. Independent later human receptor studies support the retained chemistry and location while this source-specific question remains open.
The initial reviewed draft contains 110 decisions: 72 ACCEPT, 13 MODIFY, 13 UNDECIDED and 12 KEEP_AS_NON_CORE, with no NEW or REMOVE. One integrated receptor-kinase core avoids duplicating catalytic substrate binding, self-association and ligand recognition as separate functions. There are 71 reference assessments, preserving all 65 seeded reference identities/titles and adding the six directly used donor/current papers. Exact source, quote, census and schema/history/render checks are recorded in the final integrity and handoff reports after independent review.
The full validator completed successfully, as did history validation and rendering. All 137 ordinary evidence excerpts match their actual cached source verbatim. The four warning groups are retained transparently: the six missing publication caches; the deliberately contextual generic NCK2 interaction; source-specific broad-versus-receptor kinase refinements; and a GO:0043235 label-version disagreement. The current official AmiGO term page names the term “signaling receptor complex” and lists “receptor complex” as a synonym, whereas the validator expects the latter. The verified current label and all machine-seeded fields are retained. Independent review covered all 110 decision reasons and the integrated core, with a separate molecular-function peer reading the receptor-interface and fusion-scope issues. Publication packaging remains on hold for the six required cache recoveries; the unread linked 1997 erratum remains a separate source-access question.
The six required records were recovered by the normal hosted source23 fetcher and imported byte-for-byte by the root after separate archive, identity and scientific-access checks (tmp/source23-alk-import-receipt.json). This closes the six repository-cache gates described above; those earlier failure entries remain a chronological record. All 21 explicitly used PMID records and all 42 Reactome records are now present. The unread linked 1997 erratum remains a separate unresolved bibliographic/access question, with no invented identifier or claim about its effect.
Actual local access is four abstracts (PMID:11121404, PMID:17487225, PMID:30497772, PMID:32442405) and two body extractions (PMID:23667670, PMID:41771741). Primary identities/titles/DOIs were independently checked against indexed PubMed and original publisher pages; direct PubMed opens sometimes met a browser challenge. Hosted normal retrieval, primary-web access and earlier local DNS failures are separate routes. External original-body readings for the Cell thinness study and indexed Neuron figures remain explicitly distinct from the newly cached abstracts. In particular, the Neuron abstract does not name ALK and cannot itself substantiate its Alk-specific knockdown experiment.
The zebrafish full body confirms the source-specific developmental interpretation: cloned full-length zebrafish alk and controlled morpholino/transgenic experiments distinguish mitotic/progenitor measures from neuronal differentiation, survival and positioning. Reduced proneural expression and increased apoptosis do not establish their causal order. The mouse oral-sensory study separates developmental Alk deletion from adult ceritinib exposure. Its Discussion says ceritinib also inhibits INSR/IGF1R and allows those targets to contribute to the stronger drug phenotype. Thus the drug-induced loss of neurons is not treated as the knockout phenotype or as a clean ALK-only perturbation. The recovered body contains inconsistent dose units in Methods and Results; no dosing protocol is reproduced or silently repaired here.
All 110 annotation actions, all machine-sourced objects and the integrated core remain unchanged after these source reads. Reference access flags and assessments now describe actual cached material; the developmental rationale includes the inhibitor caveat. No new annotation or additional citation was manufactured. The final integrity report separately records source preservation, the complete authored/provider census, exact cached quotations, validation and rendering.
The exact published baseline for PR #3316 is deb567575c08a06311054ec840a74f3f148440fc; all five gene files and the published history were byte-checked before authoring. The full review was read. This session preserves all 110 original annotation source objects, all reference assessments, raw files and prior history. It changes the replacement destination of the two ALKAL interaction rows and clarifies four existing uncertain reviews; it adds no NEW annotation.
The two IPI rows for human ALKAL2/Q6UX46 and ALKAL1/Q6UXT8 now refine generic binding to GO:0019955 cytokine binding. Re-reading the actual cached PMID:26418745 Results shows purified human FAM150A binding the human ALK extracellular domain by ELISA and surface plasmon resonance, plus reciprocal immunoprecipitations for both FAM150A and FAM150B. The Fc fusion in the binding preparation is an assay tag. The binding-specific replacement preserves the measured interaction and the named source partner; replacing it with phosphotransfer duplicated a different molecular function. PMID:34646012 independently establishes the human cytokine-binding interface. No claim of equal ligand potency is added. Live GO:0019955 has protein binding as its parent. GO:0004896 describes cytokine binding coupled to transmembrane signaling; the narrower binding description best matches these IPI measurements while the integrated receptor kinase core already describes signal coupling.
The three GO:0030298 rows remain UNDECIDED for a mechanistic reason, despite full source access. Live GO:0030298 and its GO:0140677/GO:0098772 ancestry require direct noncovalent activation. The 2018 paper measures activating ligand fragments, whereas the 2021 paper additionally measures genuine receptor–receptor contacts and loss of ligand-driven proliferation with the expressed M751T interface variant. These positive receptor experiments cannot be dismissed because ALK is a receptor. They also do not isolate allosteric stimulation of a partner kinase from positioning that permits trans-phosphorylation. An interface-controlled experiment fixing receptor phosphorylation state while measuring independent partner substrate turnover would resolve that distinction. Each row now states this question directly; uncertainty is not falsely attributed to unavailable text. A categorical over-annotation verdict would assume that all receptor-contact effects are covalent, which the inspected experiments do not establish.
For the source-limited GO:0032991 complex row, a conditional refinement to GO:0043235 is now explicit if its own complex proves to be the signaling receptor assembly. The separate receptor-complex result cannot identify an unrecovered ALK-specific assembly from PMID:23382219, so the original uncertainty remains. Live GO:0043235 names signaling receptor complex, with receptor complex as a synonym; the current authored label and original machine label therefore remain unchanged despite the older local validation snapshot.
The direct heparin-binding assertion remains ACCEPT. PMID:25605972 reports receptor binding and chain-length-dependent activation, and PMID:34646012 contains an additional heparin/ectodomain assembly experiment. Core prose now makes its experimental context explicit. The single compact core represents the integrated receptor kinase mechanism; it does not enumerate every component binding activity. The optional contributes_to_molecular_function slot is not used merely to fit an additional term, because the receptor itself binds heparin. The positive row and its original evidence remain available for machine-readable binding. Repeated source reasons elsewhere were not globally rewritten in this bounded follow-up; their source-specific host and construct limits remain intact.