Human BARD1 is UniProt Q99728, HGNC:952, NCBI Gene580. The source record has four alternative products; their identifiers, names and sequence-difference descriptors are preserved. The full-length protein and engineered experimental fragments must not be treated as functional measurements of every alternative product.
PMID:34321665(https://pubmed.ncbi.nlm.nih.gov/34321665/) provides structural, binding and ubiquitination evidence using recombinant human sequences. BARD1's ankyrin/BRCT region recognizes histones, DNA and N-terminal H2A ubiquitin marks, while its RING region helps position the ligase on the nucleosome. BRCA1 provides the observed E2 interface. Keep recognition of RNF168-generated H2AK13/K15 ubiquitin distinct from the heterodimer's modification of C-terminal H2A K125/K127/K129. Constructs include engineered fusions and isolated domains as well as a near-full-length BARD1 construct. The publisher's author correction adds a citation and related sentence; it does not retract the assays.
The first core function combines the demonstrated ubiquitin-modified histone reader activity with a contribution to histone H2A ubiquitin ligase activity. The existing K127/K129 assertions remain in their original source rows. H4K20me0 recognition is described without adding redundant broad and narrow reader annotations. The BRCA1–BARD1 complex belongs in in_complex; nucleus is the anatomical location.
PMID:12890688(https://pubmed.ncbi.nlm.nih.gov/12890688/) supports K6-linked autoubiquitination in the tested heterodimer system. PMID:17873885(https://pubmed.ncbi.nlm.nih.gov/17873885/) shows E2-dependent differences in output, so K6 is not asserted as the exclusive linkage for every substrate. PMID:15184379(https://pubmed.ncbi.nlm.nih.gov/15184379/) reports NPM1 stabilization after ubiquitination; ubiquitination is not equated universally with degradation.
PMID:28976962(https://pubmed.ncbi.nlm.nih.gov/28976962/) shows that purified BRCA1–BARD1 binds DNA and RAD51 and promotes synaptic complex formation and D-loop formation. BARD1 RAD51-contact mutations impair stimulation while retaining DNA binding, and complementary BRCA1 mutants also compromise activity. This supports a complex-contributed recombinase activator function, distinct from RAD51's strand-exchange catalysis and BRCA2-like presynaptic loading. The independent annotation consultation agrees with one new molecular-function assertion for this activity; no new biological-process assertion is needed because homologous recombination is already represented.
PMID:39261729(https://pubmed.ncbi.nlm.nih.gov/39261729/) separately reconstitutes stimulation of DNA-end resection machinery. EXO1 and DNA2 perform nuclease reactions, and BLM/WRN supply helicase activity. The BRCA1–BARD1 contribution is not grounds for assigning those catalytic activities to BARD1. The source's DSB-end resection assays do not establish the narrower replication-fork resection assertion in GO:0110025.
The original XIST paper PMID:12419249(https://pubmed.ncbi.nlm.nih.gov/12419249/) includes anti-BARD1 precipitation experiments despite its BRCA1-focused title. The available original Results support association without resolving direct RNA binding. Later disagreement about BRCA1-dependent XIST localization does not automatically refute that BARD1 association. Retain the curated RNA-binding observation as non-core.
PMID:10477523(https://pubmed.ncbi.nlm.nih.gov/10477523/) reports CstF-50 binding and inhibition of polyadenylation; PMID:15905410(https://pubmed.ncbi.nlm.nih.gov/15905410/) links the heterodimer to damage-dependent RNAP II ubiquitination/degradation and reduced 3′ processing. These provide specific regulatory contributions, without making BARD1 the RNA-cleavage or polyadenylation enzyme. The apoptosis and cell-cycle experiments are context-dependent; the proapoptotic p53/Ku70 mechanism and antiapoptotic BRCA1 nuclear-retention experiment need not have the same outcome.
Two original assertions remain UNDECIDED. For PMID:35512704, the exact BARD1–AKT1 wild-type/variant constructs were not resolved from the inspected material. For PMID:29709199, selected MRN-review passages do not establish BARD1-specific replication-fork resection. Neither uncertainty is framed as evidence that the curator cited the wrong gene or that the original experiment did not exist.
All 53 Reactome-derived annotations are nucleoplasm locations. Event titles describing phosphorylation, nucleolysis or another repair reaction are contextual information, not evidence that BARD1 executes every reaction. Disease-event wording is not a NOT qualifier. The original PAINT node assertions are retained; a short donor list or BARD1 itself among donors is not circular evidence. No complete family phylogeny was available, and the unreviewed family prose was not used as biological proof. No BARD1/Q99728 entry was found in the local GO-CAM index.
The standard Falcon/perplexity-lite invocation failed before either provider started because its pinned client dependency was unavailable offline. No provider report was manufactured. The review uses the normal source caches and explicitly bounded primary-paper readings. Full text flags were checked against actual content: several records remain abstract-only, and no complete supplementary-data or figure-image audit is claimed. Exact source and alternative-product objects are preserved; only authored curation fields are changed.
The integrated review preserves 161 source assertions and four alternative products, and adds one complex-contributed molecular function. The decisions are 96 ACCEPT, 50 KEEP_AS_NON_CORE, 13 MODIFY, two UNDECIDED and one NEW. The two additional papers were recovered as exact normal records after the ordinary cache command failed with DNS errors; both canonical records were verified equal to the authenticated staging bytes before application. No downloaded source text was edited.
Focused validation passed on 2026-09-30 with 33 advisories for supported generic interactions retained as non-core under the supplied ActionEnum. Rendering passed, and a matching Codex history record was scaffolded with just new-history. The two unresolved annotations remain explicit. This is a focused-gene result; the earlier repository-wide reference check remains incomplete because of an unrelated Crossref DNS failure.
Four survey-derived BRCA1 interaction annotations are refined to protein heterodimerization activity (GO:0046982), using independent primary experiments in PMID:10026184 and PMID:11573085. The former characterizes preferential heterodimer formation by RING-region constructs; the latter resolves the heterodimeric solution structure. The four original source papers and canonical BRCA1 supporting accessions remain unchanged. Their complete supplementary pair/construct records were not independently adjudicated, and proximity labeling or co-purification is not presented as a stoichiometric assay. The refinement is supported by the additional studies of the same molecular pair, consistent with the other BRCA1 refinements.
The remaining 29 supported generic-binding annotations stay non-core. The user-supplied ActionEnum reserves REMOVE for annotations unlikely to be correct and KEEP_AS_NON_CORE for supported non-core annotations. That explicit instruction takes precedence over the skill and validator recommendation to remove uninformative generic interactions. Their reasons now give biological evidence and limits without referring to a session policy. No specific catalytic, adaptor or process function is invented from a survey association alone. The unresolved AKT1 construct scope remains UNDECIDED.
Chromatin remodeling (GO:0006338) remains ACCEPT. Its current definition covers chromatin reorganization and its ontology includes histone-modifying activity as part of the process. PMID:34321665 directly measures the heterodimer's H2A modification and shows BARD1's nucleosome-positioning contribution. SMARCAD1 supplies the downstream remodeling activity discussed by the authors; BARD1 is not assigned SMARCAD1's ATPase activity. This positive direct histone-modification evidence, rather than an upstream-depletion phenotype alone, supports the broad process.
The two core locations remain nucleus (GO:0005634). The official nucleoplasm definition excludes chromosomes and the nucleolus; replacing the whole location of these chromatin-associated activities with nucleoplasm would therefore narrow their scope without necessity. All existing nucleoplasm observations remain accepted. The complex-contributed recombinase-activator NEW assertion is unchanged; its short source anchor is now also placed on that row. All occurrences of quotations, including this repetition, are counted together per source and remain at most 25 words.
Replication-fork resection (GO:0110025) remains UNDECIDED. The canonical PMID:29709199 cache contains Abstract/Introduction only despite its full-text flag. The longer recovered normal XML version was inspected separately at all BARD1 mentions and throughout the MRN replication-fork-dynamics section; it was not substituted for the canonical source. Those passages discuss BARD1 with RAD51 stimulation and discuss fork resection for MRN and its regulators. They do not identify the direct BARD1 evidence underlying the narrow NAS assertion. Direct DSB-end resection and fork protection are distinct from resection at a stalled fork, so neither is automatically substituted as a correction. The full cited experiment set and supplements remain unadjudicated.
Claim-specific reference assessments now distinguish verified paper identity from unresolved support for BARD1–AKT1 and fork resection. PMID:35512704 compares wild-type and mutant interactions; its title does not establish that every recorded edge is mutant-specific. The dispute attached to PMID:12419249 is confined to the broader BRCA1–XIST localization model, as challenged in the PMID:17146760 abstract. It does not invalidate BARD1-associated RNA observed in the original immunoprecipitation. The later paper's full Methods were not read.
This follow-up preserves all 161 source annotation objects, the one previously proposed molecular function, four alternative products, both cores and all 104 reference identities. It changes only four actions, yielding 96 ACCEPT, 46 KEEP_AS_NON_CORE, 17 MODIFY, two UNDECIDED and one NEW. No new source fetching or provider retry was needed.
The PMID:11573085 anchor now begins at “structure” so the parsed quote is a
contiguous substring of the normal abstract cache. The prior wording matched
after whitespace normalization but crossed a source newline after “solution”;
shortening it preserves the structural claim and reduces the quoted word count.
The cached publication is unchanged.
For the chromatin-remodeling annotation, the relation used in the rationale is
GO:0140993 histone modifying activity part_of GO:0006338 chromatin remodeling,
as shown by the official AmiGO term record.
This is a molecular-activity-to-process relation; it is not a claim that
GO:0016570 histone modification is an is_a child of chromatin remodeling.
Together with the directly reconstituted BRCA1–BARD1 H2A modification in
PMID:34321665, it supports the retained broad process annotation without
assigning BARD1 the ATP-dependent remodeling activity of SMARCAD1. No annotation
action, source assertion, core function, or reference record changes.