ADGRV1 (Q8WXG9) annotation review notes

Gene: ADGRV1 / GPR98 / VLGR1 / MASS1 / USH2C. HGNC:17416. Human, NCBITaxon:9606.
Largest known GPCR / cell surface protein (full-length VLGR1b ~6,307 aa).

Core biology (wild-type focus)

Photoreceptor periciliary complex (vision)

Signaling

Pleiotropic / non-core

Over-annotations flagged

Notes on evidence access

2026-09-26 — substantive re-review and source-scope corrections

This session re-examined all 64 seeded assertions, the existing structural-molecule NEW
proposal, all reference findings and all three core functions. The preceding notes are
historical: the signaling, cytoplasm, developmental-regulation, bone-mineralization and
source-access conclusions below supersede the corresponding earlier statements.

Identity, baseline and protected source material

Genuine research and cache outcomes

The existing genuine Falcon report and its artifacts were preserved. A new default
Falcon attempt with a 1,200-second timeout and the configured perplexity-lite fallback
ran concurrently with just fetch-gene-pmids human ADGRV1. Both research routes failed
before provider execution because uvx could not resolve PyPI while obtaining
deep-research-client[cyberian]==0.2.7rc1 (exit 2; approximately 9.6 and 8.2 seconds).
No replacement provider report was created. The cache sweep found all 12 existing PMIDs
already cached. A normal just fetch-pmid attempt for PMID:24962568, PMID:35630584,
PMID:36139365, PMID:24191038 and PMID:22419726 failed DNS for every item (0/5 cached,
exit 1). These primary sources were independently read through web routes as described
below, but no publication file was fabricated or modified. The gene remains DRAFT
pending these five caches. Logs are /tmp/ADGRV1-research-attempt.log,
/tmp/ADGRV1-publication-cache.log and /tmp/ADGRV1-additional-cache.log in this session.

Receptor signaling and cleavage: distinguish the actual constructs

Cytoplasmic and sensory compartments

Structural participation, developmental scope and evidence gaps

Verification checkpoint

Annotation source objects and existing reference identifiers/titles were preserved exactly.
The author is the assigned annotation-reviewer consultation; the coordinator will independently
read the stable draft before publication. Targeted schema, ontology, quotation, history and
render checks are recorded in the session history and publication manifest when complete.
The five missing caches remain an explicit draft gate regardless of validation warnings.

Independent coordinator review read all 65 decisions, three cores and reference judgments.
The signaling core uses the generic GPCR signaling BP to cover the demonstrated coupling
contexts; the Gi-specific pathway remains the explicit replacement on the two inhibitor-MF
rows. PMID:23382219 retains full_text_unavailable: false because substantial full-text
sections are available; missing target Results/SI are recorded separately and still require
UNDECIDED. The coordinator found no other biological blocker.

Final targeted validation passed with the five missing-cache warnings grouped together and
an advisory that no annotation directly cites the existing Falcon report. Standalone schema,
all cached title/quotation checks, annotation-source integrity, history validation and HTML
rendering passed. Final actions: 41 ACCEPT, 11 KEEP_AS_NON_CORE, 8 MODIFY, 3 UNDECIDED,
1 MARK_AS_OVER_ANNOTATED and the retained structural NEW. The five missing caches remain
the draft gate; final file hashes and validation logs are in /tmp/ADGRV1-audit-manifest.json.

2026-09-27 — CI reference-title metadata correction

The CI reference fetch for PMID:36139365 recovered the machine title with
Ca(2+), whereas the visible PubMed/publisher title renders a superscript.
Matched the YAML title exactly to the machine-fetched title reported by
CI run 36282024981; this changes neither the identifier nor any biological
assessment. PubMed confirms the same article and DOI (10.3390/cells11182790).
Normal local fetching still fails DNS, so the five required cache gaps remain
draft gates. CI also reported publisher-PDF HTTP 403 responses; external PDF
reading recorded above is separate from successful normal cache creation.

2026-09-27 — PR #3192 evidence-verifiability follow-up

The complete review comment
was read against published head 676504ebc6d7997336569b87b00642f26cb61230.
Local YAML, HTML and notes byte hashes matched that publication receipt before
editing. All 65 assertion source objects, including the 64 original assertions
and retained structural NEW proposal, and all 26 current reference identifier/title
pairs are preserved. The earlier title-correction record is unchanged.

A fresh supported just fetch-pmid attempt for PMID:22419726, PMID:24191038,
PMID:24962568, PMID:35630584 and PMID:36139365 returned exit 1 and cached 0/5.
Every attempt failed with nodename nor servname provided, or not known.
The complete log is /tmp/ADGRV1-followup-fetch.log. CI's earlier success reaching
PubMed metadata justified this retry, but does not establish successful retrieval
from this local runtime. No publication file was authored or modified manually.

The four open-access sources account for 28 quotation instances: 10 each from
PMID:24962568 and PMID:35630584, five from PMID:24191038 and three from
PMID:36139365. These now use supporting_text; the nonpublic-PDF exemption is
not appropriate simply because a normal fetch fails. They are not yet
cache-validated. The missing caches remain a publication gate, and future
normal retrieval must check these excerpts against the resulting text. The
PMID:36139365 excerpt now states the observed ER/MAM fraction result rather than
only that cells were fractionated. The separately sourced, abstract-only cached
PMID:16434480 still uses the full-text field for its externally read author PDF.

Primary identity and content checks remain distinct from cache availability:

The description now states the isoform/coupling boundary compactly. The
GO:0048513 developmental replacement is GO:0060122 stereocilium organization,
matching the demonstrated structural contribution and the integrated core;
the broader existing inner-ear-development assertion remains non-core context.
The PDZ-docking core omits redundant visual perception while retaining
photoreceptor maintenance. Establishment of protein localization remains in
that core because physical tail docking supplies an anchoring contribution,
in addition to the mouse partner-mislocalization phenotype. It does not assign
ADGRV1 the myosin transport step. The cached PMID:17567809 support is attached
to the core as well as the assertion. Broad cytoplasm assertions remain at
their source resolution; fragment ER/MAM evidence supports this compartment
without making the source assertion erroneous or requiring a new localization.

The review remains DRAFT. The action counts are unchanged; all five source-cache
gaps must be resolved before the evidence-verifiability request can be closed.

The coordinator independently read the complete follow-up delta, including the
description, stereocilium replacement, GPCR/PDZ cores and all five changed source
assessments, and accepted the changes. The coordinator separately confirmed
preservation of source objects, actions and reference identities. No cache was
recovered. The standard gene validation and the resulting publication manifest
record this remaining limitation rather than treating field conversion as a
successful source fetch.

Validation outcome (2026-09-27 UTC): the existing standard just validate human ADGRV1 run completed successfully, with two warning groups: unavailable publication caches and unused Falcon evidence. The normal retry recovered no caches. All five missing PMID cache gates therefore remain open, and the review remains DRAFT. Rendering, history validation, source-object and reference-identity preservation checks passed. The coordinator independently accepted the complete biological delta.

2026-09-27 normal publication-cache recovery

All five required records are recovered. PMID:35630584 (PMC9146371) and
PMID:36139365 (PMC9496679) contain XML full text; their local
full_text_unavailable flags are now false. Relevant construct Methods and
signaling or fractionation Results agree with the previously inspected primary
articles. Human VLGR1a and engineered CTF remain distinct from intact VLGR1b;
transfected human CTF localization remains distinct from endogenous retinal
experiments in other species. Existing full-text quotations for these two papers
are present in the caches.

PMID:22419726, PMID:24191038 and PMID:24962568 are abstract-only locally,
so their flags remain true. The bone-density abstract does not newly resolve
positive mineral deposition. The two signaling papers had previously been
read in full externally, and that provenance remains explicit. Their ordinary
supporting_text snippets that came from external Results are now replaced by
exact cached abstract sentences supporting the same Gi coupling, Gs/Gq-linked
calcium response and MAG stability claims. The external construct/assay details
remain in the unchanged reasons with access provenance in the source assessments.
No full-text exception field or fabricated cache was used.

The exact cache records originate from normal fetch output in Actions run
36286975328, head 5946477c8ac79ade0709264c775ea1262b108438, artifact
10920674630. The verified ZIP SHA-256 is
c0ffe4a66b80278af34b44aab6a3ae354ffd5699236b3a486ca95527be5e9713;
tmp/verified-reference-records/local-import-receipt.json records per-file
hashes. Only these five gene-required caches enter the closure manifest.

All 65 original source assertions, annotation actions and reasons, core
functions and their biological descriptions, 26 reference identities,
machine/provider files and previous history are preserved. Only evidence
attachments, source-access notes and status metadata are updated. This entry
supersedes earlier missing-cache status. Targeted validation, rendering and
history checks determine the final status and are recorded in the manifest;
any remaining advisory is distinguished from the now-closed cache gate.

2026-09-27 sensory-process follow-up at the reviewed cache-complete head

All seven canonical files matched reviewed head
c34c4a5f7eb1051687be087abc31e854c2ff86fc before editing. This follow-up
reassesses the four visual/light-perception rows named by the reviewer and
the additional ARBA light-perception row with the same core-forwarding
rationale. All five are retained as KEEP_AS_NON_CORE with source-specific
positive reasons. Their original evidence, references and qualifiers are
unchanged; none is removed or labeled as a curator error.

The live visual-perception definition
and light-perception definition
describe stimulus reception, signal conversion and neurological recognition.
The photoreceptor-maintenance definition
describes prevention of photoreceptor degeneration and is part of retina
homeostasis. These are different process scopes; maintenance is not assumed
to be a formal subtype of sensory perception. QuickGO's combined endpoint
failed, but the stated definitions were independently recovered from the
primary ontology/MOD pages on 2026-09-27.

The full genetic source PMID:14740321 establishes ADGRV1/VLGR1-related
USH2C. The cached clinical abstract PMID:15671307 reports three USH2C
siblings studied by perimetry, ERG and OCT, with rod/cone dysfunction and
outer nuclear layer thinning. The latter is positive visual-function
phenotyping, not an experiment isolating receptor-dependent phototransduction.
Human PDZ-domain interaction evidence and zebrafish connecting-cilium
localization/retinal survival data PMID:20440071 support the periciliary
complex mechanism. Thus the synthesized retinal core remains photoreceptor
maintenance, while the wider visual outcome is retained as non-core
physiological involvement. This distinction is based on the demonstrated
molecular work and source resolution, not simply on whether a GO label is
literally repeated in the core table. It does not establish that ADGRV1
cannot have a direct light-evoked signaling role.

The IBA retains the ancestral assertion with no inferred node-placement
error; the ISS retains mouse-ortholog involvement and the original
acts_upstream_of_or_within qualifier; the ARBA internals remain unverified
while independent human observations support the target-level judgment.
Each human IMP retains its own clinical evidence and access scope. The
broader hearing annotations remain ACCEPT because the ankle-link protein
supplies physical bundle cohesion with measured directional transduction
phenotypes PMID:17567809, beyond a generic disease association. The
core lists GO:0050910 once; the redundant separate GO:0007605 entry was
removed because the primary MGI ontology
explicitly places GO:0050910 part_of sensory perception of sound. This
core simplification does not change any original sound-perception action.

GO:0045184 remains broad establishment of protein localization: physical
PDZ docking plus loss of partner positioning supports the existing
source-resolution judgment. No narrower cargo-recognition or motor term
is newly inferred. All 64 seeded source assertions plus the pre-existing
structural-molecule NEW proposal, all 26 reference identities/assessments,
three molecular core units, machine/provider artifacts, exact publication
caches and published histories remain preserved. Only the five stated
review actions/reasons and the redundant core process listing change.
No new citation or cache is needed. Validation, render, new append-only
history and exact file hashes accompany the follow-up manifest.

2026-09-27 — Post-merge source-12 and recursive DOI audit

The seven gene files and five published histories exactly matched main
fb53b36b6141d82016b012f1d45c6d1dc9901541 before this follow-up. The full final
PR #3192 review and comment were read. Its merged biological judgments remain
supported: all 64 seeded assertions, the retained structural NEW proposal,
all 65 actions, the three core units and the original 26 reference identities
are preserved. No new annotation is proposed. Published history, UniProt,
GOA and both genuine Falcon artifacts remain unchanged.

The earlier 17-PMID completion statements omitted DOI-only provider sources.
This entry supersedes those statements about source completeness, rather than
rewriting the historical receipts. Nine such journal records were recovered
normally by source 12 and imported with exact byte verification; the receipt
is tmp/source12-canonical-import-receipt.json. Five contain article bodies
and four are locally abstract-only. Availability below refers to the actual
local extraction; primary identity verification and external reading are
recorded separately.

Recovered source Local access and source-specific assessment
PMID:23180093 Abstract-only review, DOI 10.1007/s12031-012-9911-5. Hair-cell architecture and earlier mouse studies provide context; this is not a new signaling assay.
PMID:33851099 Full Results/Discussion, DOI 10.1016/j.isci.2021.102283. Human RPE1 perturbations and separate mutant-mouse astrocyte experiments support focal-adhesion and mechanosensing roles. The cached Methods refer construct details to a supplement. N-terminal release as an activation mechanism remains proposed.
PMID:34331125 Abstract-only USH2 review, DOI 10.1007/s00439-021-02324-w; complex and clinical context, not an additional ADGRV1-specific experiment.
PMID:35353227 Full review, DOI 10.1007/s00439-022-02448-7. Classification and PDZD7/digenic discussion remain secondary synthesis with the authors' qualifications.
PMID:37002809 Abstract-only locally, DOI 10.1111/bcpt.13869; external full Methods/Results read at the Wiley primary article. Human CTF affinity and cellular LC3/p62 experiments support contextual regulation, without defining an autophagic catalytic or cargo-receptor step.
PMID:37127773 Full mouse study, DOI 10.1038/s41598-023-34361-y. Adgrv1 loss redistributes AC6; kinase/function tests concern Adcy6 knockout. These are distinct from directly testing receptor–cyclase coupling.
PMID:37371069 Full zebrafish study, DOI 10.3390/cells12121598. Frameshift, periciliary localization, partner-expression/localization and ERG findings support retinal maintenance with model-specific limits.
PMID:37422204 Abstract-only clinical cohort, DOI 10.1016/j.ajo.2023.06.026; 30 patients from 28 families, not a biochemical receptor assay.
PMID:37893031 Full two-sister case study, DOI 10.3390/biomedicines11102657. Variant modeling does not experimentally establish altered Calx calcium binding or digenic causation.

The focal-adhesion and AC6 papers were also independently read by the peer
annotation reviewer. The former supplies positive cellular association and
perturbation evidence, with TAP construct species unresolved from the cached
supplement pointer; it does not newly demonstrate G-protein coupling. The
latter's HEK293 expression of mouse G-alpha constructs serves antibody
validation. Its inferred ADGRV1/G-alpha/AC6 pathway is not a direct receptor
activation experiment. These distinctions preserve the existing core and
source-specific uncertainty rather than treating cell host or pathway
membership as proof of a human full-length receptor assay.

Preprints, later journal records and the two remaining cache gates

Three DOI-only preprint works remain explicit in the recursive census.
They are not silently equated with later journal versions or discarded
because no exact preprint PMID was established.

  1. 10.1101/2024.04.25.591120 is the astrocyte preprint retained in the
    Falcon report. Its author-uploaded original
    was read. Mouse astrocyte glutamate handling is regulatory context, not
    ADGRV1-catalyzed transport. The related later journal article is
    PMID:42002803, DOI 10.1186/s40478-026-02282-2, titled The adhesion GPCR
    ADGRV1 controls glutamate homeostasis in hippocampal astrocytes supporting
    neurons.     Primary PubMed identity
    and full publisher Methods/Results
    were inspected. Human Q8WXG9-1 CTF affinity assays use HEK293T and mouse
    lysates; mouse astrocyte uptake/coculture assays and one human patient
    fibroblast transcriptome have separate scopes. Association with GLAST
    complexes is not a purified binary interaction or intrinsic glutamate
    transport by ADGRV1.
  2. 10.1101/2024.09.10.612265 is the retinal transcript preprint in the
    provider report. Its full original body was not recovered in this
    follow-up. The related journal record PMID:40037841, DOI
    10.1101/gr.280060.124, is Deciphering the largest disease-associated
    transcript isoforms in the human neural retina with advanced long-read
    sequencing approaches.     PubMed identity
    and indexed primary PMC12047242 Results
    were read; direct article retrieval was challenge-limited. Human retinal
    sequencing adds transcript detail but does not reconstruct complete
    VLGR1b. The journal's exon coverage is 3–77 and 80–90, distinct from
    the earlier preprint coverage quoted by the provider. These versions
    must remain distinct; no new protein isoform identifier is invented.
  3. 10.64898/2026.03.05.709805 (the v1 URL normalizes to the same work)
    is the 2026 inactive-state structural preprint described above. Its
    prior author-copy reading and preliminary construct-specific scope are
    preserved. No exact PMID was established; it remains a bounded research
    lead, not a replacement for the established coupling evidence.

Both later journal papers are substantively used to assess retained provider
claims, so their missing normal caches are genuine review gates. The one
ordinary CLI attempt for PMID:40037841 and PMID:42002803 finished with exit 1,
0/2 cached and DNS errors; no record was manufactured. Exact terminal output
and the receipt are in tmp/ADGRV1-source12-followup/new-journal-fetch.log
and normal-fetch-receipt.json. Fresh GraphQL main-path checks found both
records absent. They are reserved for the next finite recovery proposal;
existing recovery inventories were not expanded.

Public evidence-field and rendering correction

The two public author-PDF excerpts for PMID:16434480 no longer use
supporting_text_fulltext, whose scope is private or unshareable material.
Their source remains the original author PDF,
Methods p. 762 and Figures 4–5: “PDZ1 domain interacted with VLGR1b” and
“Whirlin co-localized with the cytoplasmic tails of USH2A and VLGR1b”.
The human fragment/whirlin and COS-1 host scope remains as documented above.
These are primary-paper excerpts, not independent evidence from these notes.
No self-referential notes citation replaces the paper. Existing cached
abstract evidence and the original PMID/reference assessment are preserved.
This supersedes the earlier entry's full-text-field usage. Native PMID link
markup is used to avoid doubled links in the rendered page.

The final recursive source census contains 28 distinct PMIDs: 26 have normal
caches and the two later journal records remain missing. It also retains the
three separately identified preprints. All nine newly included cache files
were absent from the checked main tree and are listed explicitly in the
publication manifest. The review stays DRAFT. Validation, rendering,
append-only history and byte/source-preservation results accompany that
manifest; source availability is not inferred from a passing validator.

Final checks passed: gene validation (the two missing records and the existing
unused-provider advisory), history validation, rendering, 82 exact cached
quotation checks, immutable-source preservation and rendered-link inspection.
The coordinator independently read the full semantic delta and selected
primary passages and found no biological blocker. Actual unchanged action
counts are 36 ACCEPT, 16 KEEP_AS_NON_CORE, 8 MODIFY, 3 UNDECIDED,
1 MARK_AS_OVER_ANNOTATED and 1 retained NEW. The two cache gates remain open.

2026-09-27: source20 closure and direct localization-source attachment

This follow-up starts from published PR #3304 head
2a38d57ecadfb6ee4cfd65936f5963c827405b14. The current formal review
5330605982 and full comment 5856526332 were read independently. All 65
annotation source objects and actions, the three core functions, four
alternative products and 37 reference identities are preserved.

Source recovery and assay scope

Source20 recovered the two previously missing journal records through the
normal publication fetcher. The strict archive review verified exact bytes,
source-code and request/provenance pins, logs and raw copies. The coordinator
imported the records exclusively after independent primary identity/body
assessment. No cached source or provider artifact was edited.

Source Actual recovered access and scientific scope
PMID:40037841 Full XML Methods, ADGRV1 Results, Figure 6 and Discussion were inspected. PacBio analysis uses three postmortem human neural-retina samples; ONT uses three independent human retina samples. VLGR1a transcript coverage is complete, while VLGR1b remains incompletely captured. Journal coverage spans exons 3–77 and 80–90, with no exon-77-to-80 junction establishing deletion of exons 78–79. These are transcript observations, not proof of extra functional protein isoforms. The earlier preprint remains separately identified.
PMID:42002803 Full XML Methods, Results and Figures 5–9 were inspected and the live publisher version was checked. The physiological experiments use Adgrv1/del7TM mouse hippocampi and primary astrocytes/neurons. RFP-Trap uses human Q8WXG9-1 CTF residues 5891–6306 in HEK293T lysates mixed with mouse astrocyte or hippocampal lysates. One 57-year-old USH2C patient's fibroblast transcriptome, with no reported epilepsy, supplies separate human evidence. GLAST association and Stachel/randomized-peptide/receptor-deficient controls support receptor regulation of glutamate handling, without assigning ADGRV1 transporter chemistry or assuming equivalent human astrocyte physiology.

Primary identities were independently checked through PubMed, including the
PMID/DOI/PMCID/title relationships. The full astrocyte article is available
at the primary publisher.
The retinal journal article is indexed at PMC12047242;
its direct browser page was challenge-limited in this session, while its
normal XML cache contained the inspected primary sections. Both local
full_text_unavailable flags are now false. Correct citation identity was
already VERIFIED; missing local cache availability did not invalidate it.

Direct Figure 5 evidence and the autophagy boundary

For the cytoplasm IDA originally attributed to PMID:16434480, the unrelated
2022 MAM-fraction quote was removed and the original 2006 primary reference
was restored as the supporting source. The public author PDF
was independently reread: Methods p. 762 specify the human constructs;
Figure 5, p. 757, shows the 150-residue VLGR1b cytoplasmic tail in COS-1
cells, including its nucleus/cytoplasm distribution alone, cytoplasmic
retention with whirlin and loss of that retention after PDZ-binding-motif
deletion. The exact short caption excerpt already recorded above is now
attached directly in the annotation reason with the original URL/page.
The local abstract-only cache remains unchanged. No assumption about a
private or nonshareable full text is made, and these notes are a receipt
for the primary source, not an independent experiment. The retained
KEEP_AS_NON_CORE decision remains specific to the fragment-expression assay.

For PMID:37002809, upstream receptor signaling can regulate autophagy
without the receptor being an autophagic enzyme or cargo receptor. The
reference assessment now states that explicitly and preserves the positive
human-cell flux/perturbation evidence and separate mouse transcriptomic
scope. Deferring an additional NEW annotation in this bounded closure does
not deny the observed regulatory phenotype; its precise receptor-dependent
step remains a biological question. The local record is still abstract-only,
and the previously documented external full-source access is not relabeled
as a recovered local body.

The recursive authored/provider citation census is rechecked against the
unchanged reports, including DOI and URL citations. Both source20 cache
gates are now closed; no new paper was requested and no preprint was
silently remapped to a journal publication. Validation and final integrity
results are recorded below after the checks finish.

Final checks: full gene validation passed with the single pre-existing
unused-provider advisory; the YAML stays DRAFT for that intentional advisory,
while the PR source gates are closed. The complete recursive census is 28
cached PMIDs and 18 DOI identities, including three explicitly distinct
preprints. All 81 ordinary cached quotations pass case-sensitive,
whitespace-normalized substring checks; none relies on case folding. The
external Figure 5 receipt remains clearly attributed to its primary PDF.
Source preservation, published-history hashes, append-only notes, rendered
PMID links, YAML anchors and trailing whitespace were checked. The newly
scaffolded history and final rendering are validated in the handoff receipt.

2026-09-27: Figure 5 evidence and redistribution scope

The original author PDF for PMID:16434480 is publicly readable but carries
an explicit Oxford University Press rights notice on page 751. Its full text
is therefore not copied into the publication cache. This is the schema's
intended distinction for supporting_text_fulltext: access to a public URL
does not establish permission to redistribute the whole article. The earlier
public/private interpretation above was too narrow.

The Figure 5 caption on page 757 was independently rechecked at the
author PDF.
Its short verified excerpt is restored to the cytoplasm row's machine-readable
evidence field. The reason retains the human VLGR1b fragment, COS-1 host,
whirlin-dependent retention and page/URL context. This restores the contextual
literature-support check without changing the non-core localization decision
or implying that the intact receptor is freely cytosolic. The ordinary local
cache remains abstract-only, and no full article or invented cache is added.