This review distinguishes BPTF's own molecular activity from the activity of the NURF complex, preserves all 45 seeded source objects and the three named products, and records unresolved experiments explicitly. All source term IDs, labels, references, evidence, partners and annotation flags remain unchanged. No NEW annotation is proposed. The current draft contains 21 ACCEPT, 13 KEEP_AS_NON_CORE, two MODIFY and nine UNDECIDED decisions.
BPTF is a regulatory subunit of NURF. Human purification experiments identify it with SNF2L and RbAP46/48. Tagged-BPTF purification and reciprocal complex isolation support membership and co-association; they do not prove that every subunit pair forms a purified binary interaction. The primary paper measures ATP hydrolysis and remodeling for the complex, while separate BPTF depletion and promoter-occupancy experiments support Engrailed regulation. The accessible neurite-outgrowth manipulation is SNF2L. PMID:14609955(https://pubmed.ncbi.nlm.nih.gov/14609955/)
The source29 contributes_to annotation is preserved. Its proposed refinement to ATP-dependent chromatin remodeler activity describes the complex-level remodeling to which BPTF contributes; it does not assign ATPase catalysis to BPTF. The own molecular function is histone reader activity, supported by refinement of source5 rather than a redundant new annotation. The core combines this reader function, contribution to remodeling, NURF membership, nuclear location and transcriptional regulation as one mechanism. Official GO definitions and their parents were inspected: histone reader activity, ATP-dependent chromatin remodeler activity.
The 2011 study distinguishes peptide affinity from nucleosome recognition. Its bromodomain binds several acetylated H4 peptides; in the tested mononucleosomes, coupled PHD–bromodomain recognition favors H3K4me3 with H4K16ac. Full-length BPTF mutant ChIP at HOXA9 in HEK293 cells links the recognition module to chromatin recruitment or stabilization. Mutations in the pockets and their relative arrangement weaken localization. This supports reader/adaptor function beyond isolated histone affinity. It does not establish a universal preference across all nucleosome substrates or assign an assay peptide to a unique histone gene. PMID:21596426(https://pubmed.ncbi.nlm.nih.gov/21596426/)
Later primary work provides useful context without becoming fabricated cached support. The 2024 study tests a human C-terminal reader construct and reports H3K4me3 combined with H3K14/H3K18 acetylation; it did not test the older combined H3K4me3/H4K16ac condition, so the results are not a direct refutation. Its enrichment/EC50 values are not silently treated as equilibrium dissociation constants. The 2026 study's accessible captions identify human PHD2 assays and structural work on combined H3K4me3 and H3K9AcS10p recognition; complete complex recruitment and in-vivo species assignments remain outside the inspected boundary. These papers are web-only contextual notes, not cached references or quote sources in the YAML. PMID:38319148(https://pubmed.ncbi.nlm.nih.gov/38319148/), PMID:42165139(https://pubmed.ncbi.nlm.nih.gov/42165139/)
Human erythroid BPTF ChIP/sequential ChIP supports promoter chromatin occupancy. The genetic knockout described in the inspected associated experiments is Setd1a, not Bptf; BPTF is not a histone methyltransferase. The nuclear-damage study's longer authenticated supplemental Results explicitly include BPTF among recruited bromodomain proteins despite its ZMYND8-centered title. This supports nuclear localization without transferring ZMYND8's particular NuRD/repair mechanism. PMID:27141965(https://pubmed.ncbi.nlm.nih.gov/27141965/), PMID:25593309(https://pubmed.ncbi.nlm.nih.gov/25593309/)
The FAC1 studies establish sequence-selective DNA binding, sensitivity of that binding to phosphorylation, and transcriptional reporter repression. The MAZ study describes recombinant association and interaction-dependent regulation in NIH3T3 cells. The KEAP1 study describes cytoplasmic colocalization and perinuclear redistribution in the tested cell context; association does not confer KEAP1 enzymatic activity on FAC1. These are retained as supported context-specific activities and locations. PMID:10403843(https://pubmed.ncbi.nlm.nih.gov/10403843/), PMID:10575013(https://pubmed.ncbi.nlm.nih.gov/10575013/), PMID:10727212(https://pubmed.ncbi.nlm.nih.gov/10727212/), PMID:15379550(https://pubmed.ncbi.nlm.nih.gov/15379550/)
Official figure text identifies GST-FAC1 residues 1–398 in the older binding assay. The NURF paper distinguishes a FAC1 construct from full-length BPTF; its caption numbering is not used to manufacture a current amino-acid/product equivalence. UniProt associates the original FAC1 sequence study with a region of BPTF and records sequence cautions. The current seed contains exactly products 1 (Q12830-1), 2 (Q12830-2, VSP_020402) and 4 (Q12830-4, VSP_020405). Their identities and sequence notes remain unchanged. No product 3 is invented and no legacy construct is assigned to one of these products without exact mapping.
Human developmental-brain and spinal-cord studies describe changing FAC1 nuclear/cytoplasmic/cell-body distributions. Localization in ALS tissue is not interpreted as a causal ALS mechanism. The transferred cell-body and perinuclear annotations are corroborated by relevant human observations; the exact rat-source dendritic and nerve-growth-factor-response experiments remain unresolved. Unspecified neurites are not automatically dendrites. PMID:7621746(https://pubmed.ncbi.nlm.nih.gov/7621746/), PMID:9225734(https://pubmed.ncbi.nlm.nih.gov/9225734/)
The two supported NURF co-association rows remain KEEP_AS_NON_CORE under the standing ClinGen curation instructions, published in tracker161 at commit 0f1fa8fc763a6e6d05ef73a0ad5a8f8bcc7ec1ba. This is the documented task-specific departure from informational-exclusion REMOVE for a correct but generic interaction. It does not certify unavailable source-specific evidence or change global curation policy.
The four binding rows whose comparative assay or screen entry remains unread are UNDECIDED: the CERF comparison, epidermal differentiation study, and two OpenCell partner records. Separately supported BPTF/NURF biology is retained; title emphasis on another complex, or absence of a BPTF token in searchable body text, is not evidence of a wrong-gene annotation. The specific screen rows, assay scores and controls have not been independently audited. PMID:15640247(https://pubmed.ncbi.nlm.nih.gov/15640247/), PMID:22729083(https://pubmed.ncbi.nlm.nih.gov/22729083/), PMID:35271311(https://pubmed.ncbi.nlm.nih.gov/35271311/)
The two brain-development annotations remain unresolved at their exact BPTF perturbation/developmental-endpoint boundary; the NURF/Engrailed evidence is not erased and the SNF2L neurite manipulation is not silently recast as BPTF. The exosome HDA entry remains unresolved pending its BPTF-specific peptide/table evidence. Detection would not itself establish a secretion mechanism. PMID:14609955(https://pubmed.ncbi.nlm.nih.gov/14609955/), PMID:18570454(https://pubmed.ncbi.nlm.nih.gov/18570454/)
PAINT assertions are assessed as phylogenetic judgments, not pairwise donor counts. BPTF's occurrence among descendant evidence is expected rather than circular. No new PAINT tree/alignment reconstruction is claimed. Nuclear/NURF annotations supported independently by direct human evidence are retained with explicit curator deference where an original image or proteomic table remains unread.
All 15 cached publication abstracts and bibliographic metadata were read. Selected Results/Methods were read for the two new XML full-text papers, including the full-length BPTF mutant-ChIP passage in the 2011 reader study. Longer authenticated variants of two already-existing caches were used only for bounded supplemental reading; canonical caches are not overwritten or normalized. Direct PMC openings that returned challenges are recorded, and accessible PubMed captions/indexed primary passages are distinguished from full article, figure-image and supplement inspection. No complete-paper audit is claimed for the full-text caches merely because their metadata flag is true.
Nine brief exact anchors are allocated across the draft, with no source exceeding 25 quoted words. The provider attempt, canonical import closure and normal schema/term/reference/GOA checks remain pending at this draft stage. This note will receive a dated status entry after those checks; no application, publication or completion is claimed here.
The three normal primary seed files and eleven new publication caches have now been imported from the authenticated Seed63 artifact and independently confirmed. The four existing publication-cache variants were preserved byte-for-byte, the two PANTHER family exports remain quarantined, and all 22 recent campaign source caches were preserved. The candidate still contains the same 45 source annotation objects, three named products, 23 references, nine short excerpts and one synthesized core. No NEW annotation or source normalization was introduced.
One standard Falcon deep-research invocation with the configured Perplexity-lite fallback was attempted after source import. Both routes stopped during offline tool-dependency resolution for deep-research-client[cyberian]==0.2.7rc1. Neither provider produced a research report. These are infrastructure failures, not evidence about BPTF biology. The manual review relies on the source reading and access boundaries documented above; no manual text is presented as provider-generated output.
The normal focused validation with ontology terms passed with no errors. Three warnings remain explicit: two supported generic NURF-association rows retained under the task instruction, and the distinction between general BPTF/NURF transcriptional regulation and a historical FAC1 construct-specific regulation experiment. Independent strict GOA validation also passed. The core-synthesis consultation found the single reader/remodeler-contribution unit coherent, with no intrinsic BPTF ATPase assignment. These are candidate checks; canonical curation application, rendering, session history and publication remain pending independent scientific peer review.
The curation-instruction link above points to the exact published tracker commit. Its presence on that tracker branch is not a claim that a relative project link is already available on every gene branch. Publication must carry a reachable instruction reference or consume the independently checked shared instruction when it is available on the gene branch.
The independently reviewed candidate is now applied to the canonical BPTF review and notes. Normal canonical validation passed with the same three explained warnings: two supported generic NURF associations and the distinction between general transcriptional regulation and a historical FAC1 construct experiment. The standard session history was scaffolded and validated, and the targeted rendered page presents this final local state. All 45 original source objects, three products, 15 publication-cache bytes, the additional recent-source protections and the two quarantined family exports remain unchanged. Earlier statements that application was pending are historical. No GitHub approval, publication, merge or campaign completion is claimed. The task instruction remains linked to its exact published tracker commit; this does not assert that the shared policy is already on main.
This subsection records the proposed current interpretation; earlier draft counts and unresolved-action statements above remain historical. The proposal preserves all 45 source annotation objects, their partner/qualifier fields and all three named products. It introduces no NEW annotation. Canonical application and publication have not occurred for this follow-up.
The original H4-partner binding object is refined to histone H4K16ac reader activity. A separately supported histone H3K4me3 reader activity represents the PHD arm in the core synthesis. These sibling terms replace the generic parent core; they describe coupled arms of one NURF recruitment mechanism, not independent pathways. Remodeling remains a contribution to the complex, never intrinsic BPTF ATPase activity. Definitions and parent paths were checked against the official GO release dated 2026-03-25; unsuccessful live API access is not represented as an October refresh.
The 2011 primary study, PMC3135172 provides peptide, reconstituted-nucleosome and full-length BPTF mutant-ChIP evidence at HOXA9. Its Figure 6B links the paired recognition module to recruitment; genome-wide mapping used the reader cassette. H4K16ac preference depends on the nucleosome context, whereas isolated acetyl-H4 peptides show broader binding. The 2006 H3-recognition paper provides structural and affinity evidence. The companion primary paper, also available as an author-hosted Nature PDF, tests human BPTF recognition and chromatin recruitment; its Xenopus knockdown/rescue is a separately identified developmental experiment. The 2007 structural paper changes methylation-state preference with an engineered PHD mutation, not a new wild-type activity inferred from L3MBTL1.
The immutable UniProt INTERACTION section separately records Q12830-4 contacts with H4C9/P62805 and H3C15/Q71DI3. This resolves those database product assignments, but does not establish that every historical reader assay used that product. The original GOA objects are not rewritten to add an unprovided isoform identifier; legacy FAC1 construct mapping remains distinct.
The 2017 ISWI study, PMC5623870 supports endogenous BPTF association with both SMARCA1 and SMARCA5 through ATPase immunoprecipitation and reciprocal BPTF antibody precipitation. These are cellular coassociation experiments, with nuclease treatment and ATPase depletion controls. Residual nucleic acid and trace alternative ATPase are acknowledged by the authors; direct binary BPTF–SMARCA5 contact is not established by these assays. The recombinant double-affinity purification panel expressly excluded the 340-kDa BPTF protein, and the listed expression constructs also omit BPTF. Consequently, its purified direct-interaction, ATPase and nucleosome-sliding measurements cannot be assigned to an intact recombinant BPTF/NURF-5 complex. The normal HTML cache contains Table 1's heading and footnotes but not its cells. NURF-1 and NURF-5 composition/naming here is explicitly attributed to the immutable reviewed UniProt SUBUNIT record, not an inspected primary table. The UniProt statement comparing their ATP hydrolysis rates is not independently established by this recombinant panel and is not adopted in the biological summary.
The 2004 SNF2L splice-variant paper shows in its cached abstract that the inactive SMARCA1/SNF2L+13 variant retains incorporation into multiprotein complexes. That abstract does not name BPTF. UniProt and the 2017 paper attribute stable BPTF association to this earlier study, but its original BPTF-specific assay remains uninspected. The +13 splice event concerns the SMARCA1 motor, not a BPTF isoform. These findings support separating motor competence from complex assembly without implying that reader activity guarantees remodeling in every associated complex.
The four pairwise binding objects previously left UNDECIDED are proposed as KEEP_AS_NON_CORE with independent same-partner corroboration and explicit curator deference. Original CERF-comparison, epidermal-study and OpenCell assay details remain uninspected; none is marked independently verified on the strength of another paper. Unlike a multi-partner aggregate with unknown additional edges, each object asserts the single preserved SMARCA1 or SMARCA5 partner that is corroborated. No ATPase, chaperone or catalytic activity is transferred from the partner.
For the existing brain-development annotations, BPTF promoter occupancy and reduced en-1 expression after BPTF depletion provide a mechanistic connection, while reviewed UniProt explicitly records the developmental role. The proposed non-core retention defers to the existing curator process interpretation; it does not claim an independently inspected BPTF whole-brain phenotype. The original paper's Figure 7 distinguishes BPTF depletion from SNF2L depletion. Its neurite-assay Methods, PMC275440 use SNF2L constructs. The unchanged positive-transcription annotation concerns a measured transcriptional endpoint; that distinction is retained.
NEDDFL (MIM:617755) provides relevant disease context. The human haploinsufficiency study reports developmental/speech delay and frequent postnatal microcephaly, with separately identified zebrafish loss-of-function experiments. Its abstract and captions were read, not a complete clinical or experimental audit. This supports a dosage-mechanism question, not additional GO process annotations. The exosome HDA, rat-source dendrite and nerve-growth-factor-response objects remain UNDECIDED; nuclear identity alone does not refute an experimental exosome identification.
Previously inspected supplemental material is publicly identifiable: human NURF Figures 2, 3 and 7, FAC1 binding captions, and BPTF damage-recruitment Results, PMC4298138. Their URLs make the evidence traceable without pretending that the shorter canonical caches contain the inspected additional passages. No source cache is expanded or normalized, and no supplemental-only excerpt becomes a canonical supporting quote.
The user's standing ClinGen task instruction is to retain supported, biologically correct generic protein-binding annotations as KEEP_AS_NON_CORE when no evidence-backed specific replacement is established. This intentionally differs from the repository skill's informational-exclusion REMOVE recommendation: lack of specificity does not mean the interaction is false. Unresolved biological evidence remains UNDECIDED. This paragraph records the task instruction, not a new global policy or an external maintainer sign-off. The instruction is now in the project page at the independently confirmed main merge. This immutable link is usable from the gene branch independently of its base. The follow-up publication should incorporate the policy-bearing main under fresh source-preserving branch guards; these proposed notes do not claim that such a branch update has already occurred.
Reference notes are shortened to findings plus the particular access limit. All nine existing source quotations remain verbatim. The five additional landmark references now have verified normal caches after the single ordinary fetch failed DNS and the reviewed source-recovery route completed. Their exact imported titles, DOI/PMCID values and availability flags are retained: 15310751, 16728976 and 16728978 are abstract-only; 18042461 and 28801535 contain HTML-extracted text. A PMCID does not turn 16728978 into a full-text cache. The new source-specific excerpts share each paper's aggregate 25-word budget, including excerpts used in core support and reference findings. The final proposal has 28 references (20 PMIDs), 21 ACCEPT, 19 KEEP_AS_NON_CORE, two MODIFY and three UNDECIDED decisions, with two specific reader core entries and no redundant parent core. All 45 machine source objects and three products remain exact. This is a proposal awaiting distinct science review; no follow-up gene application or publication is asserted.
The independent science review passed and the exact follow-up YAML and notes were applied. Normal canonical gene validation and the standard scaffolded history validation passed. The documented eight warnings preserve the task binding instruction, FAC1 assay context and independently supported H3 reader synthesis. Source objects, named products, existing histories and source-cache bytes remain unchanged. Rendering and remote publication are recorded separately; this local application is not a merge or external approval claim.
The review now proposes one NEW IDA annotation for GO:0140002 histone H3K4me3 reader activity, in addition to the 45 unchanged machine-sourced annotation objects. The preceding journal entries describe earlier review cuts; their no-NEW statements and action tallies are historical, not the current proposal. The current tally is 21 ACCEPT, 19 KEEP_AS_NON_CORE, two MODIFY, three UNDECIDED and one NEW (46 objects total). All 47 raw GOA rows, their existing 45-object projection, three named products and two coupled reader cores remain unchanged.
The new assertion uses PMID:16728978(https://pubmed.ncbi.nlm.nih.gov/16728978/) for direct structural and binding evidence on the human BPTF PHD finger. The normal cache is abstract-only, and no new complete-paper, image or supplementary-data inspection is claimed here. PMID:16728976(https://pubmed.ncbi.nlm.nih.gov/16728976/) links H3K4me3 recognition to BPTF chromatin association and NURF motor recruitment; its Xenopus developmental experiments are kept separate from the human molecular-function claim. The already reviewed PMID:21596426(https://pmc.ncbi.nlm.nih.gov/articles/PMC3135172/) supplies coupled-reader and full-length BPTF recruitment evidence. These findings support a reader/adaptor role, not BPTF ATP hydrolysis. No historical construct is assigned retrospectively to a current product, and H3K4me3 preference is not asserted as complete exclusion of H3K4me2.
GO:0140002 is absent from the existing source and replacement terms. It is a sibling of the H4K16ac reader term, not an ancestor or descendant of that replacement. None of the seven original generic-binding rows has an H3 source partner, so no unrelated partner record is repurposed. The one new four-word exact cached anchor shares PMID:16728978's quotation budget with the unchanged core excerpt (17 words combined). All 15 earlier supporting excerpts remain exact and unchanged.
The reference inventory is 28 records: 20 PMIDs, seven GO_REF method records and one UniProt file reference. The review comment's total of 21 omitted the seven GO_REF records; none is removed. The ClinGen curation instructions on merged main provide the durable policy link for this follow-up. Earlier dated notes retain their historical tracker link. The instruction is the user's scoped task direction, not a global skill change or external maintainer approval.
The additional annotation is a curation proposal; its inclusion does not assert publication, approval or merge. Independent review, application and validation are recorded separately. The three unresolved exosome, dendrite and nerve-growth-factor evidence judgments remain unresolved.