UniProt Q9UL18 (AGO1_HUMAN, EIF2C1). Taxon NCBITaxon:9606.
Created as a 2026 update to an already-complete review. Scope is one topic: the disputed
AGO1x stop-codon-readthrough proteoform, which the review did not mention (the file
contained zero occurrences of "readthrough", "AGO1x" or "stop codon" before this update).
No. This was checked first, before writing anything, because it determines whether an
edit is warranted at all.
AGO1-goa.tsv cites 36 PMIDs plus GO_REFs and Reactome ids. None ofAGO1-uniprot.txt has no ALTERNATIVE PRODUCTS block, no isoform records, and no mentionSo the readthrough dispute does not affect the annotation set. No action changed on any
of the 113 existing annotations. Saying so plainly is the deliverable here; manufacturing an
edit would have been the wrong outcome.
alternative_products was deliberately not used: that slot maps 1:1 to UniProt isoform
IDs and no UniProt isoform for AGO1x exists, so any entry would have required inventing an
identifier. functional_isoforms was also not used — its type enum (SPLICE_VARIANT,
SPLICE_CLASS, CLEAVAGE_PRODUCT, MODIFICATION_STATE, CONFORMATIONAL_STATE) has no category for
a translational-readthrough proteoform, and creating an entry would assert the proteoform
exists, which is precisely what is disputed. The dispute is recorded in knowledge_gaps,
references, suggested_questions and description instead.
PMID:31330067 — i.e. AGO1x is proposed as a
competitive inhibitor of the miRNA pathway, generated by let-7a-promoted readthrough of the
AGO1 stop codon.
An independent laboratory reported the same proteoform with a different function:
PMID:32812257, localizing
atypically PMID:32812257.
So AGO1x is not a single-group claim.
A Letter from the Atkins/Loughran group, who specialize in readthrough measurement, argues
the quantification was a reporter artefact:
- With an insulated StopGo/2A single-mRNA reporter: PMID:40500330, while the AQP4 positive control gave the expected ~6%.
- Rebuilding the original fused AGO1-F-luc construct reproduced apparent readthrough, which
vanished on removing the fusion: PMID:40500330.
- Mechanism of the artefact: an 8-fold depressed fused sense-codon control plus
reporter-independent background, so PMID:40500330.
- Why it was suspicious in the first place: PMID:40500330
The Letter's title is about the dual luciferase assay, and that is genuinely its scope. It
names PMID:31330067 and PMID:38499809 and addresses their reporter measurements. It does
not address:
- the orthogonal evidence in the same 2019 paper — PMID:31330067 and PMID:31330067;
- the independent 2020 work, whose antibody specificity was controlled against FLAG-AGO1 and
FLAG-AGO1x lines PMID:32812257.
So the honest reading is: the reported readthrough efficiency is very likely wrong; whether
the proteoform exists at some lower level is open. No reply from the originating
laboratories was indexed in PubMed at the time of this review (searches: "AGO1x",
"AGO1 stop codon readthrough", "AGO1 readthrough reply").
The two published AGO1x functions are opposite in sign or orthogonal to canonical AGO1: a
GW182-incompetent miRNA-pathway inhibitor, and a nuclear dsRNA/interferon suppressor. AGO1
carries several non-core nuclear annotations. Should AGO1x turn out to be real, some
nuclear observations attributed to AGO1 may belong to the extended proteoform; should it
not, two functional models rest on an artefact. Recorded as a knowledge_gaps entry rather
than acted on.
Two pre-existing validator warnings remain (IBA annotations at index 5 and 7 marked
MARK_AS_OVER_ANNOTATED without structured propagation_review). These predate this update
and are unrelated to it; the PAINT trees were not inspected, and per project rules
propagation metadata must not be fabricated to silence a warning.
This section supersedes the older narrowly scoped review above where it says no reply
existed, calls the reported readthrough rate probably wrong without discussing the
response, or says the PAINT review remains unperformed. The older text is retained as
session history, not the current scientific conclusion.
Identity and baseline: human AGO1, UniProt Q9UL18, HGNC 3262; former symbol
EIF2C1, alias hAGO1. Current HGNC-derived identity was checked against
NCBI Gene 26523 and
ClinGen.
The parent independently checked current main
795b693f5711c625401d03a755fe937260ae0ac0: all five review/notes/HTML/UniProt/GOA
blobs matched the local baseline, and separate AGO1 and EIF2C1 open-PR searches were
empty. The seeded review contains 113 grouped annotation objects. Every original
source object, including its NOT flag and any qualifier/isoform field, is preserved.
The 82 original reference identifiers and titles are also preserved. No NEW annotation
was added, and no machine source was edited.
A fresh genuine Falcon request with perplexity-lite fallback was launched concurrently
with publication caching, using the normal research wrapper and task-specific writable
UV tool/cache directories. Both provider commands failed before research submission:
Failed to fetch: https://pypi.org/simple/deep-research-client/, DNS resolution failure,
provider exit code 2; the wrapper returned 1. No new provider report exists. The genuine
May 2026 Falcon report already present in the gene directory remains unchanged. It was
read as a lead, including its explicit limitation on direct AGO1 nuclear evidence, and
was replaced as the sole support for the annotation decisions by source-specific
primary evidence. Publication caching found all 34/34 original YAML PMIDs cached.
Logs are /tmp/AGO1-fresh-research.log and /tmp/AGO1-fetch-pmids.log.
The human protein is directly assayed rather than inferred solely from AGO2:
Both seeded NOT process annotations are retained. Live AmiGO definitions for
GO:0035279 and
GO:0090625 explicitly describe
small-RNA-directed target cleavage, despite the “mRNA destabilization” labels and,
for the former, deadenylation synonyms. The original abstract PMID:15260970 directly
compares human Argonaute complexes and identifies AGO2-specific cleavage. These NOT
rows must not be read as absence of AGO1-recruited nonendonucleolytic decay. A proposed
expert question requests definition/name/synonym harmonization; the source terms and
NOT flags themselves are untouched. The full 2004 paper was not recovered, so no
additional assay detail beyond the original abstract is claimed.
The precise RISC-loading complex assertions remain UNDECIDED. The current
GO:0070578 definition describes a
DICER1–TARBP2–AGO2 trimer. The full PMID:19966796 explicitly questions whether this
historical human RLC is the canonical loading intermediate. Human AGO1 pre-RISC and
AGO1–Dicer association PMID:12526743 do not alone establish the exact trimeric
composition. This uncertainty does not deny AGO1 loading/assembly.
The eight IBA rows were traced to their actual IBDs in
interpro/panther/PTHR22891/PTHR22891-paint.tsv:
PTN001875625 (nucleus), PTN000527275 (cytoplasm), PTN001113179
(RISC, cytoplasmic RNP granule, miRNA binding and silencing), PTN000527278
(pre-miRNA processing), and PTN000527276 (ssRNA binding). Structured sources use the
PTN ancestors. Human AGO1 among extant evidence seeds is legitimate direct grounding,
not circularity. The PAINT placement and human primary evidence agree; the tree/MSA
was not newly reconstructed. ARBA rule identities and combined InterPro/UniProt/Ensembl
branches were recovered from WITH/FROM, but uninspected rule predicates are explicitly
UNRESOLVED even where independent human biology supports ACCEPT.
The Ensembl donor is mouse Ago1 Q8CJG1, ENSMUSP00000095498.
Final source tracing recovered the MGI graph, generated 2023-03-10 (a historical annotation snapshot, not a current experiment). Its promoter GO:0000976 IMP, GO:0000978 IDA and nucleus IDA trace J:197739 = [PMID:22053081, Upregulation of Cyclin B1 by miRNA and its implications in cancer.]. The existing full XML cache was read. Endogenous Ago1 knockdowns are in mouse cells; tagged human AGO1 is explicitly used in NIH/3T3 ChIP and compartment-fractionation experiments. Promoter recruitment and nuclear localization are positive results. The discussion considers promoter-derived noncoding RNA targets, so precise DNA contact/sequence recognition remains unresolved; nuclear location is independently supported.
The exact GO:1901224 IMP traces J:233964 = [PMID:26764146, Argonaute proteins in cardiac tissue contribute to the heart injury during viral myocarditis.] (primary PubMed abstract). This resolves donor identity. The abstract reports AGO1/AGO3 upregulation in CVB3-infected mouse hearts and in-vitro repression of miR-19a/b-targeted TNFAIP3, an NF-kappaB inhibitor. It does not expose the NIK/non-canonical branch assay. Publisher full text returned 403; normal fetch-pmid 26764146 failed DNS, cached 0/1, and produced no file. Retain the exact branch UNDECIDED without asserting that an unobserved assay was absent. MouseMine search snippets joined adjacent publication titles, so the paper title was verified directly against PubMed before citation.
Full PMID:25336585 independently provides human AGO1 ChIP at the IL2 promoter,
AGO1-dependent let-7i reporter activation, and HA-AGO1 co-immunoprecipitation with Pol II
and TBP. These support contextual nuclear localization, transcriptional activation and
Pol II-complex binding. The promoter/guide compensatory sequence mutations principally
test let-7i recognition. Exact AGO1 sequence-specific DNA recognition remains unresolved
because promoter occupancy could involve the bound RNA and other protein contacts;
this is not rejection merely for lacking a purified-protein assay. The Reactome
nucleoplasmic event R-HSA-5578742 independently describes AGO1-containing chromatin
complexes. These canonical AGO1 observations do not depend on AGO1x.
Broad cytoplasmic/cytosolic annotations are retained at their actual source resolution.
All 40 cited Reactome summaries were read, including the transcription-context events
that generate miRNAs. The review distinguishes the supported AGO1 location from a
claim that AGO1 is the named transcription factor or was specifically tested in every
miRNA-target experiment. R-HSA-9925159 contains a miR-140/miR-34a inconsistency in its
reporter sentence; this does not establish a false AGO1 cytosolic compartment and is
flagged in reference assessment. HPA and electronic locations have independent human
AGO1 imaging/mRNP corroboration. The mouse-ZAR1L-titled PMID:20014101 explicitly
mentions EIF2C1/AGO1 colocalization, so its title is not used to invent species
misattribution.
All 21 generic protein-binding rows retain REMOVE under the project specificity policy,
with source-specific partner/assay explanations. The interactions are not declared
false. GW182, PRNP, importin, chaperone-client, phosphatase-screen, centrosome and
APOBEC3 contexts are distinguished. Known Ago2-focused titles that actually include
AGO1 are explicitly recognized. Partial proteomic caches do not certify every
supplementary pair record; no specific new MF is manufactured from an interaction alone.
A co-published response exists: [PMID:40500329, Response to Suresh et al.]
(PubMed,
publisher full text).
It contests the reporter design and interpretation and emphasizes orthogonal evidence.
The two letters leave the quantitative dispute open. The prior “no reply” statement
and categorical readthrough-rate conclusion are superseded. Neither letter is a
retraction. The response also disputes the critique's highest-known-human-efficiency
statement; the review no longer adopts that claim as fact.
The source findings, reference assessments, description, knowledge gap and questions
were harmonized. The independent PMID:32812257 paper contains readthrough-specific
peptide spectra, antibody controls and genomic perturbation in breast-cancer cells;
it is not simply treated as an unverified identifier because another reporter assay
is disputed. PMID:31330067 and PMID:38499809 report orthogonal assays alongside
reporters. Their disputed reporter interpretation is scoped separately from all
proteoform observations. No GOA source object depends on any of these papers, and no
invented UniProt isoform identifier was added for AGO1x.
A normal fetch-pmid 40500329 attempt failed DNS resolution and cached 0/1; no
publication file was fabricated. The new reference retains
full_text_unavailable: true because no local cache exists, despite external full-text
access. All other availability flags reflect local cache metadata. HTML extractions
for broad proteomics studies contain body sections but are partial; their false
unavailability flags are not a claim that all figures/tables were recovered.
Peer consultation: the annotation-reviewer peer independently checked
PMID:18771919/PMID:15260970, the live cleavage-process definitions,
PMID:19966796 RLC scope and PMID:25336585 promoter/Pol II assays. The peer agreed
with the bounded NOT, decay, RLC and promoter-recognition judgments. Parent is
independently auditing the complete stable draft. No source objects changed during
these consultations.
Final checks: all 113 original annotation source objects and all 82 original
reference identifiers/titles are preserved. UniProt and GOA byte hashes match the
verified main baseline. Final actions are 79 ACCEPT, 7 KEEP_AS_NON_CORE, 6 UNDECIDED
and 21 REMOVE; no NEW annotations were introduced. The integrated core and 85
reference assessments were checked for consistent scope. The notes-inclusive census
contains 38 PMIDs, with 36 cached and PMID:26764146 / PMID:40500329 missing.
Both normal fetch attempts failed DNS; this is a draft/publication gate.
The parent independently read all 113 decisions, the core and reference assessments
and found no biological blocker, then received the final bounded donor-provenance
update for inspection. just validate human AGO1 passes with only the two reported
warning categories: missing reference caches and no annotation quote from the
available provider report. The latter is intentional: annotations use checked primary
sources. just validate-history and just render human AGO1 pass. YAML is alias-free
and has no trailing whitespace. No source caches, provider artifacts, Git state or
shared project files were edited.
Changed the authored review status from COMPLETE to DRAFT to match the schema and the publication gate already documented for missing PMID:26764146 and PMID:40500329 caches. All 113 annotation decisions, source fields, reference assessments and the integrated core are unchanged. The YAML's only parsed change is the status value. The published baseline is e47a3fa018bcdc23d4a26f5ff367771f8385b5c1; its three gene-file blobs were independently confirmed by the coordinator and matched locally before editing. Targeted validation, regenerated rendering and a new append-only session record accompany this correction. The external cache requirement remains open; this status change makes no new biological claim.
The published baseline is 5aaf1b1ec8cf9b1383f85350d799148945185bf2;
the coordinator verified all three authored file blobs before editing. This section
supersedes the earlier generic-binding tally and confident acceptance of the two
negated process annotations. All 113 annotation source objects, 85 reference
identities, and original negation flags remain unchanged.
For PMID:18771919, the GO:0035278 support now quotes the translation-specific
comparison, rather than reusing the mRNA-abundance result. The full cached Results
separate similar reductions in cytoplasmic mRNA concentration from differing protein
repression: Ago4 > Ago3 > Ago2 >= Ago1. At low concentrations, Ago4 was approximately
sixfold more potent than Ago1. The paper attributes the difference mainly to
translation. This engineered 293T/boxB reporter comparison supports AGO1 activity,
but not equal paralog potency or a universal quantitative balance between repression
and decay in endogenous cells. The core and reference assessment now state that scope.
For PMID:25336585, Figure 1F identifies TBP as the AGO1 co-immunoprecipitation
partner. Generic protein binding is therefore MODIFY to GO:0017025, TBP-class
protein binding, at the same association-assay resolution as the retained Pol II
binding row. The current AmiGO definition
and MGI ontology entry
were independently checked; a direct QuickGO endpoint attempt failed. This does not
assert a purified binary contact or make the contextual nuclear function core.
Live AmiGO checks retain target-cleavage definitions for
GO:0035279 and
GO:0090625; GO:0035279 also
retains deadenylation synonyms. The original PMID:15260970 abstract and independent
AGO1 structural/cleavage work PMID:23809764 support a cleavage-scoped negative.
They do not show absence of nonendonucleolytic decay recruitment. The original
publisher full text still returned 403 and its PDF failed to load. Both process NOT
rows are now UNDECIDED pending intended ontology scope and original assay details,
without changing the source negation or the accepted NOT RNA-endonuclease MF.
Read-only comparison found AGO4's explicit scope uncertainty relevant; AGO3's miRNA
decision rests on later AGO3-specific cleavage experiments and is not evidence for
human AGO1 slicing. No paralog files were edited.
The cytoplasm and P-body core locations are retained to describe the wider
cytoplasmic pool and an observed condensate-associated pool, rather than confining
all AGO1 function to P-bodies. PMID:26764146 remains VERIFIED for the previously
checked primary abstract, citation identity and broad mechanism; inaccessible
branch-specific experiments remain explicitly unresolved. Missing local cache
does not erase that documented source verification. The two missing required
caches remain PMID:26764146 and PMID:40500329, so status stays DRAFT. No cache,
provider report, GOA, UniProt, Git state or shared project file was modified.
Final follow-up checks: 77 ACCEPT, 7 KEEP_AS_NON_CORE, 8 UNDECIDED, 20 REMOVE and 1 MODIFY. The coordinator independently read the full delta and checked the two primary assay caches and live ontology terms, finding no biological blocker. Gene validation passes with the same missing-cache and unused-provider-quote warnings; history validation and rendering pass. All source objects/reference identities and immutable source bytes match the published baseline.
Both required records, PMID:26764146 and PMID:40500329, now exist as exact normal
fetch output. PMID:26764146 remains abstract-only. Its AGO1/AGO3, TNFAIP3 and
miR-19a/b results do not identify the precise NIK/non-canonical NF-kappaB assay,
so the already reviewed mouse-donor transfer remains UNDECIDED. The known
historical MGI provenance is preserved. PMID:40500329 now provides XML full text
from PMC12264181, matching the previously read publisher reply. The HBB reporter
comparison tests stop-go reporter sensitivity; it does not independently measure
endogenous AGO1 abundance. The prior AGO1 orthogonal evidence and the opposing
methodological critique remain explicitly disputed. Its local access flag is
now false; scientific correctness remains DISPUTED.
These are unedited records from normal fetch Actions run 36286975328, head
5946477c8ac79ade0709264c775ea1262b108438, artifact 10920674630. Verified ZIP
SHA-256: c0ffe4a66b80278af34b44aab6a3ae354ffd5699236b3a486ca95527be5e9713.
Per-file hashes are in tmp/verified-reference-records/local-import-receipt.json.
This dated entry supersedes the earlier missing-cache status. All 113 original
source assertions and reviews, the core synthesis, 85 reference identities,
NOT/isoform fields, machine/provider files and prior history are preserved.
Validation, rendering and history checks are recorded in the closure manifest.
COMPLETE requires zero validation warnings; source-specific uncertainty and any
unused-provider advisory are separate from the now-closed cache requirements.
At reviewed head db1f96f0bbead7b746ad4d1ba71720bec6d9aea2, the
GO:1901224 annotation reason retained a historical fetch-failure sentence
although its reference assessment correctly described the recovered abstract.
That sentence now states the current limit: the normal PMID:26764146 cache
contains the abstract, while the full experimental body is unavailable.
UNDECIDED is unchanged because the abstract does not identify the exact
NIK/non-canonical branch assay. All 113 source objects/actions, the core,
85 reference identities and the source-specific donor assessment are unchanged.
No provider, machine, cache or published history was edited.
The recovered records' integrity can be checked directly from this published
note, independently of the local import-receipt path:
26231660ee5a88eab7359e14e1967e561b7f73ec1c00f7e3776e6ba7b846168c.a18cef782151c917d87276f6e4e6608fc71e824f6266395b64ffe36af4ad405f.The source remains normal fetch run 36286975328 at head
5946477c8ac79ade0709264c775ea1262b108438, artifact 10920674630, with the
previously recorded ZIP digest. The structured DISPUTED reference judgments
and explicit competing reporter interpretations are retained; this narrow
access correction does not choose one experimental estimate as settled.
Targeted validation, rendering and a new append-only history record accompany
this follow-up. Status remains DRAFT while validation advisories persist.
Approved AGO1 (HGNC:3262, UniProt Q9UL18; previous EIF2C1, alias hAGO1) was checked against current main 03db1457249828e300303e015f99c334168f689c. All six gene files and six published histories matched exactly; canonical and historical-symbol open-PR searches were empty. The prior functional follow-up had already addressed the signaling-reference access wording before its approved merge. This closure assesses the six additional DOI-linked provider references recovered by the normal source-12 fetcher and imported unchanged. It preserves all 113 annotation objects, their qualifiers and action decisions, the integrated core, previous reference assessments, raw sources and provider text.
| Reference | Actual evidence and limits |
|---|---|
| PMID:37066388 | Local preprint abstract. The normal record explicitly links the later publication below; the two versions are not independent replication. The tested protein is worm ALG-1 carrying equivalents of human AGO1 disease substitutions. |
| PMID:38412125 | Local main Methods, Results and Discussion. Endogenous worm alg-1 was edited by CRISPR. Developmental phenotypes, protein abundance, small-RNA association and guide/passenger ratios, RNA-seq and ribosome profiling distinguish allele-specific effects. Translational efficiency is derived from ribosome footprints relative to transcript abundance; altered steady-state mRNA levels are not direct half-life measurements. The largely recessive or weakly semidominant worm heterozygote phenotypes do not establish human dominant-negative behavior. Human residue labels and existing human AGO2 structural comparisons do not turn these into human AGO1 constructs or structural assays. Linked supplementary files were not separately read. |
| PMID:38995884 | Local clinical report: one child has de novo AGO3 and KHSRP nonsense variants. AGO1 is background comparison, not the tested locus or an original protein assay. Candidate disease association is not sufficient evidence of a new AGO1 function. |
| PMID:39457367 | Local review of rare miRNA-related diseases. Its AGO1 clinical summary, AGO2 assay discussion and proposed miRNA-turnover mechanisms are distinguished. No original AGO1 experiment is performed. |
| PMID:40863728 | Local Argonaute-interactome review. It integrates heterogeneous paralog-specific studies; chaperone ATP hydrolysis is not AGO1 ATPase activity, and generic loading-complex prose does not settle exact AGO1 RLC composition. Its AGO1x section repeats older findings and does not resolve the primary 2025 disagreement. |
| PMID:41154621 | Local miRNA/cancer review, including loading, translation/deadenylation, modifications and decay sections. AGO1-compatible nonendonucleolytic silencing is distinguished from AGO2/AGO3-specific examples. In target-directed miRNA degradation, the AGO protein is a ubiquitination/proteasome substrate; ligase and RNA-degradation chemistry are not assigned to AGO1. |
Original identities can be checked at PubMed 37066388, PubMed 38412125, PubMed 38995884, PubMed 39457367, PubMed 40863728 and PubMed 41154621. The preprint cache remains abstract-only; the other five contain the extracted bodies assessed above. Source recovery does not remove the existing source-specific uncertainties about the defined loading complex, signaling-branch assignment, promoter contact or destabilization-term semantics. The existing AGO1x critique/reply assessment is retained rather than overridden by a later secondary summary.
The protected Falcon report also cites Manuel Nuno Condinho Rodrigues Reis de Carvalho, Study of human Argonaute 1 cap-independent translation (2019), without a DOI or PMID. Primary academic records independently identify it as a Universidade de Lisboa biochemistry master's thesis: supervisor Luisa Romao's CV and examiner Ana Luisa Silva's CV. The indexed repository handle is 10451/37702. The repository record and full thesis were not retrievable during this follow-up; no exact PMID or journal replacement was established. Its full experiments remain unassessed. A study of translation of AGO1 mRNA also must not be conflated with an intrinsic translation-initiation-factor activity of the AGO1 protein. This source is retained transparently as access-limited provider context, not as an independently verified functional claim, fabricated cache or silently exempt PMID.
The recursive authored/provider census, including normalized DOI and PMC links and the title-only item above, identifies 44 required PMIDs and 40 Reactome records, all now locally present. The six new source-12 publication bytes are included in the publication manifest when absent from main. No newly missing normal-cache identifier was found, and no repeat fetch or new annotation is required. DRAFT remains appropriate for retained validation advisories; closure of download requirements does not imply that every scientific uncertainty or full-text access limitation has been resolved.