CAVIN1

Summary: CAVIN1-specific mRNA capture remains unverified.

Reason: The HeLa UV-crosslinking atlas establishes an mRNA-interactome assay, but its CAVIN1-specific identification and controls were not inspected. Conserved pre-rRNA binding provides relevant biological context without independently verifying this human mRNA-screen assertion. Retain the experimental source and uncertainty.

Summary: Hormone-sensitive lipase associates with PTRF in human adipocytes.

Reason: The human-adipocyte study reports reciprocal HSL/PTRF coimmunoprecipitation and insulin-dependent redistribution. This supports the supplied Q05469 association but does not establish a direct binding interface or CAVIN1 lipase activity. Retain the supported association as non-core.

Summary: SDPR/CAVIN2 binds and recruits PTRF.

Reason: The primary study reports cell association and binding of a recombinant PTRF 47–268 fragment to SDPR constructs, with recruitment experiments in caveolin-deficient fibroblasts. This is positive partner evidence, although the purified assay is not full-length PTRF and does not by itself specify the whole coat function.

Summary: Caveolin-3 association is supported in the caveolar system.

Reason: The human disease study links PTRF defects to caveolin mislocalization and impaired interactions; the exact P56539 pair also appears in the reviewed UniProt interaction record. Retain this experimentally curated association without converting it into direct purified binding or a muscle-specific catalytic function.

Summary: Caveolin-1 association is supported by convergent coat evidence.

Reason: The supplied Q03135 association is corroborated by the reviewed interaction record and direct analysis of the human caveolar coat. Crosslinking and co-isolation support complex membership, but the coat paper explicitly leaves the precise direct caveolin-contact mechanism open.

Summary: Cavin-2 is associated with cavin-1 coat subcomplexes.

Reason: Human-cell crosslinking, immunoisolation and partial crosslink cleavage support cavin-1/cavin-2 subcomplexes. Cavin-2 and cavin-3 occupy distinct subcomplexes within the coat. Retain the O95810 association; the broader structural function is represented by the corresponding coat-scaffold refinement.

Summary: CAVIN1 contributes the structural framework of the caveolar coat.

Reason: Human-cell coat purification, trimer analysis and electron microscopy establish CAVIN1 as a physical coat constituent, beyond a requirement inferred only from knockdown. GO:0140378 captures its own structural contribution to holding a protein complex together. The term refers to integral membership in a protein complex, not insertion as an integral membrane protein. Direct purified binding to caveolin-1 was not established.

Proposed replacements: protein complex scaffold activity

Summary: Cavin-3 occupies cavin-1-associated coat subcomplexes.

Reason: The human coat study supports Q969G5 co-isolation with cavin-1 and distinct cavin-2 versus cavin-3 subcomplexes. This preserves a supported association without attributing cavin-3-specific signaling or budding activities to CAVIN1.

Summary: CAVIN2 association is consistent with the cavin interaction network.

Reason: Selected primary Results describe cavin interactions in living cardiomyocytes, and the exact CAVIN1–O95810 pair is corroborated by reviewed UniProt and the independent human coat study. The cardiac paper uses COS cells and rodent cardiomyocytes; it is not evidence that CAVIN1 performs the CAVIN4-specific ERK-recruitment mechanism.

Summary: Caveolin-3 association is retained with assay-context limits.

Reason: The supplied P56539 pair is present in reviewed UniProt and coherent with the cavin/caveolin network studied using expressed constructs and cardiomyocytes. No CAVIN4-specific ERK or hypertrophy function is transferred to CAVIN1, and the uninspected supplementary construct details are not inferred from the cell species.

Summary: CAVIN4 and PTRF associate in the primary cellular experiments.

Reason: The accessible Results explicitly report MURC/CAVIN4 coimmunoprecipitation with PTRF in transfected COS cells and complementary interaction evidence in cardiomyocytes. This supports retention of the Q5BKX8 association. Supplementary construct provenance was inaccessible, so neither an endogenous human complex nor purified direct binding is claimed.

Summary: The CAV3 screen association is corroborated.

Reason: The exact P56539 pair is listed in reviewed UniProt and has independent caveolar-system support. Retain the BioPlex association as non-core; affinity-purification coassociation does not define the direct interface or transfer a caveolin-specific activity.

Summary: FBXO28 association is supported at the interaction level.

Reason: The exact Q9NVF7 pair is listed in the reviewed interaction record and recurs in complementary interaction screens. Target-specific biochemical function and the original AP-MS enrichment controls were not independently resolved; this supports a bounded non-core association, not an E3-ligase or substrate-regulation assignment.

Summary: The CDC37 association is retained as non-core.

Reason: The curated HuRI binary-interaction assertion for Q16543 is corroborated by the exact partner in reviewed UniProt. The target-specific original screen table was not independently re-read, and the corroborating record is not treated as a separate physiological assay. Its chaperone-cofactor identity does not establish that CAVIN1 is a co-chaperone.

Summary: The TSGA10IP association is retained as non-core.

Reason: The curated HuRI binary-interaction assertion for Q3SY00 is corroborated by the exact partner in reviewed UniProt. The target-specific original screen table was not independently re-read, and the corroborating record is not treated as a separate physiological assay. This association does not establish a germ-cell process for CAVIN1.

Summary: The SPANXN3 association is retained as non-core.

Reason: The curated HuRI binary-interaction assertion for Q5MJ09 is corroborated by the exact partner in reviewed UniProt. The target-specific original screen table was not independently re-read, and the corroborating record is not treated as a separate physiological assay. No testis-specific physiological mechanism is inferred from the partner.

Summary: The SPANXN2 association is retained as non-core.

Reason: The curated HuRI binary-interaction assertion for Q5MJ10 is corroborated by the exact partner in reviewed UniProt. The target-specific original screen table was not independently re-read, and the corroborating record is not treated as a separate physiological assay. Repeated detection does not establish a germline role for CAVIN1.

Summary: The DNTTIP2 association is retained as non-core.

Reason: The curated HuRI binary-interaction assertion for Q5QJE6 is corroborated by the exact partner in reviewed UniProt. The target-specific original screen table was not independently re-read, and the corroborating record is not treated as a separate physiological assay. Its transcription-associated identity does not establish a shared transcriptional mechanism.

Summary: The ZFP1 association is retained as non-core.

Reason: The curated HuRI binary-interaction assertion for Q6P2D0 is corroborated by the exact partner in reviewed UniProt. The target-specific original screen table was not independently re-read, and the corroborating record is not treated as a separate physiological assay. Binding to a zinc-finger protein is not evidence that CAVIN1 binds DNA.

Summary: The YAF2 isoform Q8IY57-5 association is retained as non-core.

Reason: The curated HuRI binary-interaction assertion for Q8IY57-5 is corroborated by the exact partner in reviewed UniProt. The target-specific original screen table was not independently re-read, and the corroborating record is not treated as a separate physiological assay. The partner isoform is preserved; it is not a CAVIN1 isoform annotation or proof of Polycomb regulation.

Summary: The TTC5 association is retained as non-core.

Reason: The curated HuRI binary-interaction assertion for Q8N0Z6 is corroborated by the exact partner in reviewed UniProt. The target-specific original screen table was not independently re-read, and the corroborating record is not treated as a separate physiological assay. No stress-response or acetyltransferase-cofactor activity is transferred from this partner.

Summary: The ZCCHC10 association is retained as non-core.

Reason: The curated HuRI binary-interaction assertion for Q8TBK6 is corroborated by the exact partner in reviewed UniProt. The target-specific original screen table was not independently re-read, and the corroborating record is not treated as a separate physiological assay. The association does not specify RNA substrate or zinc-binding activity for CAVIN1.

Summary: The TNFAIP8L1 association is retained as non-core.

Reason: The curated HuRI binary-interaction assertion for Q8WVP5 is corroborated by the exact partner in reviewed UniProt. The target-specific original screen table was not independently re-read, and the corroborating record is not treated as a separate physiological assay. No cell-death or lipid-transfer function is inferred solely from the partner.

Summary: The KAT5 association is retained as non-core.

Reason: The curated HuRI binary-interaction assertion for Q92993 is corroborated by the exact partner in reviewed UniProt. The target-specific original screen table was not independently re-read, and the corroborating record is not treated as a separate physiological assay. This does not establish acetyltransferase activity or a specific chromatin complex for CAVIN1.

Summary: The CAVIN3 association is retained as non-core.

Reason: The curated HuRI binary-interaction assertion for Q969G5 is corroborated by the exact partner in reviewed UniProt. The target-specific original screen table was not independently re-read, and the corroborating record is not treated as a separate physiological assay. Independent coat-complex evidence corroborates this pair, but the binary screen does not define its whole-cell role.

Summary: The CDCA7L association is retained as non-core.

Reason: The curated HuRI binary-interaction assertion for Q96GN5 is corroborated by the exact partner in reviewed UniProt. The target-specific original screen table was not independently re-read, and the corroborating record is not treated as a separate physiological assay. A cell-cycle-associated partner does not establish CAVIN1 participation in cell-cycle progression.

Summary: The MRPL45 association is retained as non-core.

Reason: The curated HuRI binary-interaction assertion for Q9BRJ2 is corroborated by the exact partner in reviewed UniProt. The target-specific original screen table was not independently re-read, and the corroborating record is not treated as a separate physiological assay. The pair does not demonstrate mitochondrial localization or mitoribosome membership of CAVIN1.

Summary: The INO80B association is retained as non-core.

Reason: The curated HuRI binary-interaction assertion for Q9C086 is corroborated by the exact partner in reviewed UniProt. The target-specific original screen table was not independently re-read, and the corroborating record is not treated as a separate physiological assay. No INO80 chromatin-remodeling activity or complex membership is inferred from this pair alone.

Summary: The FBXO28 association is retained as non-core.

Reason: The curated HuRI binary-interaction assertion for Q9NVF7 is corroborated by the exact partner in reviewed UniProt. The target-specific original screen table was not independently re-read, and the corroborating record is not treated as a separate physiological assay. This does not define an ubiquitination reaction, substrate relationship or enzymatic function.

Summary: The ZSCAN21 association is retained as non-core.

Reason: The curated HuRI binary-interaction assertion for Q9Y5A6 is corroborated by the exact partner in reviewed UniProt. The target-specific original screen table was not independently re-read, and the corroborating record is not treated as a separate physiological assay. The partner does not make CAVIN1 a sequence-specific DNA-binding transcription factor.

Summary: CAV3 coassociation is retained with AP-MS limits.

Reason: The exact P56539 pair is corroborated in reviewed UniProt, with complementary interaction evidence in the source set. The preserved canonical BioPlex cache is partial HTML; its original target-level enrichment and cell-context controls were not independently inspected. Retain a supported non-core association without inferring caveolin-specific membrane activities.

Summary: SPANXN2 coassociation is retained with AP-MS limits.

Reason: The exact Q5MJ10 pair is corroborated in reviewed UniProt, with complementary interaction evidence in the source set. The preserved canonical BioPlex cache is partial HTML; its original target-level enrichment and cell-context controls were not independently inspected. Retain a supported non-core association without inferring a germ-cell biological process.

Summary: ZCCHC10 coassociation is retained with AP-MS limits.

Reason: The exact Q8TBK6 pair is corroborated in reviewed UniProt, with complementary interaction evidence in the source set. The preserved canonical BioPlex cache is partial HTML; its original target-level enrichment and cell-context controls were not independently inspected. Retain a supported non-core association without inferring a specific RNA-binding mechanism.

Summary: CAVIN3 coassociation is retained with AP-MS limits.

Reason: The exact Q969G5 pair is corroborated in reviewed UniProt, with complementary interaction evidence in the source set. The preserved canonical BioPlex cache is partial HTML; its original target-level enrichment and cell-context controls were not independently inspected. Retain a supported non-core association without inferring CAVIN3-specific signaling.

Summary: FBXO28 coassociation is retained with AP-MS limits.

Reason: The exact Q9NVF7 pair is corroborated in reviewed UniProt, with complementary interaction evidence in the source set. The preserved canonical BioPlex cache is partial HTML; its original target-level enrichment and cell-context controls were not independently inspected. Retain a supported non-core association without inferring an ubiquitination mechanism.

Summary: DNTTIP2 association is corroborated across interaction resources.

Reason: The multimodal U2OS map supplies the curated Q5QJE6 association, and reviewed UniProt plus HuRI support the exact pair. The original target-level assembly assignment was not independently examined. Retain the interaction without assigning a transcription complex or claiming that imaging proves direct binding.

Summary: FBXO28 association is corroborated across interaction resources.

Reason: The exact Q9NVF7 pair recurs in the curated multimodal map, HuRI and reviewed UniProt. Retain the non-core association; target-specific assembly controls and a causal ubiquitination mechanism are not established by this review.

Summary: A regulated nuclear pool occurs in human adipocytes.

Reason: The human study explicitly reports insulin-dependent movement of PTRF from the plasma membrane to cytosol and nucleus. This supports nuclear localization, which is retained alongside the caveolar location. Later mouse cellular experiments establish a regulated nuclear function; the human localization evidence itself remains the cited adipocyte work.

Summary: Human adipocyte nuclei contain PTRF.

Reason: The accessible original localization Results report detection in separately purified nuclei with nuclear-marker controls. The mitochondrial/nuclear combined fraction alone would not suffice, but the additional nuclear experiment supports this annotation. Later mouse cellular experiments establish a regulated nuclear function; the human localization evidence itself remains the cited adipocyte work.

Summary: Nuclear localization is supported beyond the electronic transfer.

Reason: The combined mouse-orthology and UniProt-location mapping is corroborated by human adipocyte nuclear measurements. Retain the nuclear pool without assuming that every nuclear biochemical assay was performed on human protein. Later mouse cellular experiments establish a regulated nuclear function; the human localization evidence itself remains the cited adipocyte work.

Summary: Reactome places PTRF in the nuclear transcript-release event.

Reason: The preserved human Reactome event describes binding to and dissociation of a paused Pol I/pre-rRNA complex. It supports the pathway location as a curated nuclear role; the cached summary is not an independent human localization experiment and does not itself resolve original construct provenance.

Summary: Nucleoplasmic placement is consistent with the curated Pol I event.

Reason: The human Reactome event places PTRF with Pol I and nascent rRNA. Retain the nuclear location, corroborated at nuclear resolution in human adipocytes, while distinguishing the pathway representation from a direct nucleoplasm imaging assay.

Summary: The broad cytoplasmic location is sound but not the defining compartment.

Reason: CAVIN1 is a peripheral cytoplasmic coat component with soluble pools. The PAINT assertion represents an inherited family judgment, and inclusion of human Q6NZI2 among descendant evidence is not circular. More informative caveolar localization is retained separately.

Summary: PTRF occupies a cytoplasmic pool in human adipocytes.

Reason: The insulin-translocation study explicitly distinguishes plasma membrane, cytosolic and nuclear pools. Retain the broad cytoplasmic annotation as non-core; it does not specify the structural activity occurring at caveolae.

Summary: The original mitochondrial EXP localization remains unresolved.

Reason: The accessible human-adipocyte abstract establishes membrane, cytosolic and nuclear redistribution but does not resolve a mitochondria-specific assay. The complete mitochondrial experiment was not inspected. Preserve the experimental annotation without asserting an error or absence from mitochondria. ORGANELLES batch 2 check (2026-10-24): the paper is not in PMC and the cached abstract does not mention mitochondria, so the curator's assay is still unseen and the row stays UNDECIDED rather than being overruled.

Summary: A combined mitochondrial/nuclear fraction does not settle mitochondrial localization.

Reason: The selected original Results and fractionation caption describe low PTRF in a combined mitochondrial/nuclear fraction and independently validate nuclei. A mitochondria-specific validation was not recovered in the sections inspected. This limits verification of the IDA assertion without disproving it. ORGANELLES batch 2 check (2026-10-24): PMC1134064 returns no machine-readable body, so the fraction composition could not be re-read here; a pellet that pools mitochondria with nuclei is the crude-fraction pattern that cannot establish the organelle. Kept UNDECIDED, not overruled.

Summary: The mitochondrial location mapping needs primary clarification.

Reason: The SL-0173 mapping accurately records the reviewed location source, but the accessible fractionation evidence does not cleanly separate mitochondria from nuclei. Retain uncertainty about the mapped compartment rather than treating broad fraction enrichment as organelle-specific proof. Examined as a synthesis row (ORGANELLES batch 2, 2026-10-24): SL-0173 comes from the curated line "Mitochondrion {ECO:0000269|PubMed:15242332, ECO:0000269|PubMed:17026959}", the same two adipocyte papers as the direct rows, so it adds no independent line. CAVIN1 has no transit peptide or TM segment and is a cytosolic caveolar coat protein; the Human Protein Atlas reports plasma membrane, and a mitochondrial-ribosome partner in an interaction screen (MRPL45) is not a localization assay. The action matches the direct rows; no sub-compartment is proposed.

Summary: The endoplasmic-reticulum transfer is not yet independently resolved.

Reason: The combined annotation cites mouse CAVIN1 and the ER location vocabulary. The specific donor experiment and corresponding human localization were not inspected; a caveolar role does not rule out transient ER association. Preserve the original edge and uncertainty. ORGANELLES check (2026-11-05, ER batch 2). The source was traced. The human UniProt ER statement is by similarity to rat P85125, whose own ER line is a curator inference (ECO:0000305) beside a microsome fraction, and a microsomal fraction without an ER marker cannot carry an ER claim. The cited mouse entry O54724 tags its ER line ECO:0000269 to PMID:30188967, a Fam198a secretion paper cached as abstract only (full text not retrievable here); the abstract places the Fam198a precursor in the ER and shows it retained there in cavin-1-null cells, but does not state that cavin-1 itself was seen in the ER. CAVIN1 has no signal peptide or TM segment, so any ER pool would be a recruited peripheral pool, and the human protein is mainly at the plasma membrane. The row stays UNDECIDED because the one experimental source cannot be read; if the full text shows only an indirect role in ER export of cargo, MARK_AS_OVER_ANNOTATED is the expected outcome, and if it shows a marker-controlled ER pool of cavin-1, KEEP_AS_NON_CORE at the root. No sub-compartment is added.

Summary: Cytosolic PTRF is compatible with its soluble and peripheral membrane pools.

Reason: Human adipocyte redistribution and the coat studies support a soluble cytosolic pool. Retain the location mapping as non-core alongside the more specific caveolar coat location.

Summary: Human adipocyte PTRF is concentrated at the plasma membrane.

Reason: The vectorial proteomics and localization work identifies PTRF at the surface of caveolae in human adipocytes. This plasma-membrane association is appropriate for a peripheral coat protein and does not imply transmembrane topology.

Summary: A plasma-membrane PTRF pool is directly supported in human adipocytes.

Reason: The human study follows an insulin-responsive plasma-membrane pool and its redistribution. This is consistent with the defining caveolar coat localization.

Summary: Cavin associates with caveolin-positive plasma-membrane structures.

Reason: The primary abstract reports plasma-membrane redistribution upon caveolin expression and a role in caveola organization; independent human-cell coat localization corroborates the compartment. The exact species of every expressed construct in the earlier study is not inferred from its HEK293 host.

Summary: Human Protein Atlas imaging supports the plasma membrane.

Reason: The directly inspected CAVIN1 subcellular page reports supported plasma-membrane localization with HPA049838 and HPA074213 across human cell lines. This agrees with the primary coat studies. Vesicles on the same page are not added as a new annotation.

Summary: The plasma-membrane location mapping is well corroborated.

Reason: The SL-0039 mapping agrees with human adipocyte localization and human-cell caveolar-coat experiments. Retain it with peripheral, cytoplasmic-face association understood.

Summary: Caveolar activity is conserved and directly supported on human CAVIN1.

Reason: The PAINT family assertion is corroborated by human coat purification and electron microscopy. Human Q6NZI2 in WITH/FROM represents descendant experimental grounding, not circular evidence; no ancestral node or family label is invented.

Summary: The human adipocyte study supports caveolar localization.

Reason: PTRF is studied as a caveolar protein and associates with HSL in human adipocytes. This agrees with independent structural localization around caveolar bulbs.

Summary: Human-cell microscopy and cavin recruitment support caveola localization.

Reason: The SDPR study directly analyzes PTRF in caveolar structures and its recruitment by SDPR. The conclusion is localization of PTRF; the SDPR-specific membrane-tubulation activity is not transferred.

Summary: CAVIN1 forms part of the coat surrounding caveolar bulbs.

Reason: Human-cell immuno-electron microscopy and tagged-coat tomography place cavin on the caveolar bulb, distinct from the EHD2-enriched neck. This directly supports the specific location.

Summary: The mouse-orthology caveola mapping is independently supported in human cells.

Reason: The combined source transfer agrees with human-cell coat composition and ultrastructure. Retain the original donor and vocabulary fields exactly.

Summary: The inherited initiation assertion remains unresolved at the ancestral-function level.

Reason: GO:0006361 and its initiation parent concern promoter-associated initiation, including preinitiation-complex assembly and the first RNA bonds of a transcription round. PMID:11139612 establishes murine PTRF-dependent dissociation of paused complexes and more repeated transcription, including on promoter-independent templates; this identifies release-driven recycling rather than isolating PTRF work in an initiation step. The dated MGI donor graph links mouse initiation IDA to J:121710, and official MouseMine maps that reference to PMID:11139612. That resolves one donor paper, but does not reconstruct the PAINT IBD at PTN002689946, its other evidence or ancestral placement. The target in WITH/FROM is valid descendant evidence, not circular support. Retain UNDECIDED for the phylogenetic assertion without counting donors, declaring the IBD wrong, or excluding every possible role of recycling factors in subsequent initiation.

Summary: The original IDA source does not yet settle the initiation step.

Reason: The original 1998 biochemical Results establish paused-complex release and the Discussion leaves coupling to initiation open. Later murine reconstitution establishes recycling rather than a PTRF-dependent promoter initiation step: enhancement appears over multiple rounds and also on promoter-independent templates. This is a mechanistic distinction, not a requirement that an initiation factor increase only the first round. The human truncated clone in the original interaction screen and full-length murine protein in functional assays remain distinct; the precise human IDA mapping has not been independently resolved. Retain UNDECIDED without a wrong-gene claim or treating increased total output alone as initiation.

Summary: Electronic transfer of the donor recycling readout overextends the demonstrated initiation-step evidence.

Reason: GO_REF:0000107 transfers experimentally curated ortholog annotations, so the mouse donor is not uncurated. The official MGI graph generated in 2023 records mouse initiation IDA J:121710; the official MouseMine snapshot updated in January 2025 explicitly maps it to PMID:11139612. The inspected primary Methods and selected Results use mouse PTRF and murine templates. PTRF releases paused ternary complexes, and its increase in repeated transcription occurs even with promoter-independent templates; isolated-complex release does not increase transcripts already synthesized. Thus the experimentally resolved work is transcript/polymerase release that enables recycling, rather than demonstrated assembly of a promoter preinitiation complex or synthesis of the first RNA bonds. Transferring that readout as human CAVIN1 participation in GO:0006361 overextends the identified assay role. The 2016 mouse ChIP/3C and rescue data corroborate termination-linked output and promoter–terminator proximity, but do not isolate such an initiation step either. This bounded verdict does not claim that CAVIN1 can never contribute to initiation, that all donor curation is erroneous, or that first-round stimulation is universally necessary. Existing termination and broader rRNA-transcription assertions already represent the supported work; no redundant replacement or new process is added.

Summary: Conserved transcript release supports a distinct nuclear Pol I termination function.

Reason: The original murine reconstitution shows release of both transcript and polymerase from paused complexes. This is a substantive function consistent with the PAINT assertion and curated human Reactome events. Retain the inherited nuclear role; the IBD placement itself was not independently reconstructed. Subsequent reconstitution separates release from transcriptional recycling, and mouse adipocyte localization, TTF1 association and termination-region ChIP corroborate a cellular nuclear role. These are ortholog studies, not direct human release assays.

Summary: Human-specific provenance of the original termination IDA needs clarification.

Reason: The accessible primary biochemical assay uses full-length murine PTRF, while a truncated human clone was used for interaction screening. Conserved termination is plausible and retained through inherited/pathway evidence, but the direct experimental mapping onto human CAVIN1 was not independently resolved. No wrong-gene claim is made. This is uncertainty about the human IDA attribution, not disbelief in the conserved termination function retained through IBA and ortholog transfer.

Summary: The mouse-to-human termination transfer is biologically coherent.

Reason: The mouse donor performs transcript and polymerase release in a reconstituted paused complex, and the human Reactome representation assigns the same role to PTRF. Retain the conserved process without presenting the donor assay as a direct human experiment. Subsequent reconstitution separates release from transcriptional recycling, and mouse adipocyte localization, TTF1 association and termination-region ChIP corroborate a cellular nuclear role. These are ortholog studies, not direct human release assays.

Summary: The broad rRNA-transcription transfer is consistent with transcript-release biology.

Reason: The conserved protein has a demonstrated biochemical role in release of Pol I transcription complexes, not merely an expression correlation. Retain this broader secondary process without assigning RNA-polymerase catalytic activity to PTRF. Mouse adipocyte and MEF knockout/rescue experiments further support nutrient-responsive rRNA output, with loss of promoter–terminator proximity in a 3C assay. PTRF supplies release/regulatory work; Pol I performs RNA synthesis.

Summary: The curator-mediated rRNA-transcription transfer is coherent.

Reason: The manual mouse orthology transfer is consistent with transcript-release biochemistry and the curated human Pol I events. It is retained as a secondary function with the organism distinction explicit. Mouse adipocyte and MEF knockout/rescue experiments further support nutrient-responsive rRNA output, with loss of promoter–terminator proximity in a 3C assay. PTRF supplies release/regulatory work; Pol I performs RNA synthesis.

Summary: The protein-secretion donor assertion is not sufficiently resolved.

Reason: The specific experimental basis behind the mouse Compara transfer was not inspected. Caveola disruption can change trafficking and secretion indirectly, so necessity alone would not establish CAVIN1 doing a secretory step. Preserve the existing assertion and request donor-level clarification.

Summary: Membership in the caveolar coat complex is directly supported.

Reason: The primary human-cell study isolates the cavin/caveolin coat and visualizes its repeating units. The broad protein-containing-complex term is correct but less informative than the coat-scaffold activity and caveolar location represented in the core.

Summary: Conserved pre-rRNA recognition is supported by biochemical donor evidence.

Reason: The murine protein binds the U-rich nascent pre-rRNA region in sequence-sensitive biochemical tests. This is consistent with the PAINT inheritance assertion and the human Reactome model; inclusion of target human evidence is not circular. Retain this conserved nuclear molecular function without claiming direct human full-length binding was independently demonstrated.

Summary: The original human pre-rRNA-binding IDA provenance remains unresolved.

Reason: The inspected RNA-binding experiments used recombinant murine protein; the human fragment was obtained through interaction screening. The biochemical function is supported in the ortholog, but the human-specific experimental mapping was not independently verified. Preserve the source and uncertainty. This source-attribution uncertainty does not negate the conserved RNA-recognition function retained through IBA and ortholog transfer.

Summary: The mouse pre-rRNA-binding transfer is mechanistically supported.

Reason: Sequence-sensitive interaction with the U-rich nascent transcript provides positive donor evidence for this conserved molecular function. Retain the electronic transfer as a distinct nuclear function, while distinguishing inherited support from an independently verified human binding assay.

Summary: CAVIN1 self-association contributes to the trimeric coat unit.

Reason: Partial crosslink cleavage and immunoisolation in the human-cell coat study support cavin-1 trimers. This is a functionally informative homomeric interaction that forms part of the structural scaffold, not an unspecified screen-only association.

Summary: The self-interaction screen is consistent with independent trimer evidence.

Reason: The exact self-pair appears in the reviewed interaction record, and the primary coat study supports trimeric CAVIN1. Retain the binary-screen annotation as corroborating non-core evidence rather than using it to determine stoichiometry by itself.

Summary: CAVIN1 is recovered in human endothelial detergent-resistant membrane fractions.

Reason: The primary HUVEC proteomic table explicitly identifies Q6NZI2 and its altered fraction recovery after flotillin-1 depletion. This supports the broader membrane-raft localization, with detergent-fraction and perturbation limits; the paper does not directly establish caveolar morphology through that proteomic experiment.

Summary: The positive-motility electronic transfer needs its original donor mechanism.

Reason: The specific mouse experiment was not inspected. Changes in motility can follow altered membrane organization without demonstrating a direct motility-regulatory step. Preserve the existing donor/source fields and uncertainty rather than removing a potentially valid context-specific effect.

Summary: The manual positive-motility transfer remains unresolved at the donor level.

Reason: The orthology assertion may reflect a legitimate cellular context, but its original experiment and causal scope were not recovered. A structural caveolar function alone neither proves nor disproves positive regulation of cell motility.

Summary: CAVIN1 supplies an organizing structural unit during assembly of the caveolar coat.

Reason: GO:0070836 describes components aggregating, arranging and bonding to form caveolae, under plasma membrane raft assembly. CAVIN1 performs structural work in this process: human cavin constructs, endogenous localization controls, crosslinking and isolation identify cavin-1 oligomeric units throughout the coat, and coassembly experiments show that cavin-2 and cavin-3 enter a common complex through cavin-1. The latter tests use mouse knockout fibroblast backgrounds with human cavin constructs; human HeLa/RPE experiments independently establish coat incorporation and organization. This evidence goes beyond loss-of-caveolae necessity or CAVIN1 merely being cargo. It does not establish a purified direct CAVIN1–caveolin contact, complete atomic geometry or independent membrane-bending activity. Archived normal human CAV1 and CAV3 annotations supply positive same-coat structural comparators for this process. Dated Cavin2/Ehd2 mouse tables and inaccessible Cavin3/QuickGO results do not establish a systematic exclusion of cavins; EHD2 is a neck-associated control rather than the same coat role. No existing process annotation is an ancestor or descendant of this proposed assembly process, and the molecular scaffold and caveola location are different ontology aspects. The local GO-CAM index has no literal CAVIN1/Q6NZI2/PTRF hit; this is not an exhaustive external-model claim. Add the specific assembly process to the existing structural-scaffold core, without adding a second assembly ancestor or a membrane-bending term.