BARD1-related Cancer Predisposition: research report

Read this first. I ran four web searches in this session. Anything marked [searched] comes from those results. Mostly that is search-result abstracts and snippets, and I did not open the full texts. Anything marked [memory] is background recall with no source checked in this run. Treat it as a lead and verify it before it goes into a KB entry. I did not write any PMIDs from memory, because an unchecked PMID is a defect under this repo's rules. Where I have a source, I give the DOI or PMC ID that appeared in the results. The MONDO ID in the template (MONDO:0700267), the OMIM, Orphanet and ClinGen identifiers, and the HP/GO/CL/UBERON/NCIT term suggestions were not looked up. Do not bind any of them until they are read from the ontology.

Scope and lump/split calls

• This entry is the germline predisposition syndrome. It should stay separate from the somatic breast, ovarian and neuroblastoma entries it predisposes to (design-decisions §3a).
• Lump/split call 1. Adult breast cancer predisposition (moderate-penetrance, heterozygous) and the proposed pediatric neuroblastoma predisposition are different clinical entities with different evidence strength. I would keep one germline entry, with the two as separate phenotype arms. The pediatric arm would be flagged as lower-confidence unless a ClinGen or guideline source supports it. A split into two entries is defensible.
• Lump/split call 2. OMIM carries a neuroblastoma susceptibility locus, "Neuroblastoma, Susceptibility To, 5" [searched: findzebra listing only; OMIM number not verified]. Whether this and the breast predisposition belong in one disorder is unresolved.
• The Mendelian lane also needs an explicit decision on biallelic BARD1. I found no source on it in this session.

1. Disease information

• BARD1 encodes the obligate heterodimer partner of BRCA1. The heterodimer has E3 ubiquitin ligase activity and acts in homologous recombination repair. [memory]
• Heterozygous germline loss-of-function variants confer moderate-penetrance breast cancer risk, with an apparently stronger association with triple-negative disease. The clinical framing is "BRCA1-like partner, but lower penetrance" [memory].
• Identifiers and synonyms: not retrieved. The MONDO ID given in the template is unverified. HGNC for BARD1 is not looked up. Neuroblastoma susceptibility 5 appears in a findzebra listing [searched], but I did not verify the OMIM number.
• Data source class: aggregated cohort and case-control resources (germline sequencing series, gnomAD-controlled comparisons), not EHR-derived.

2. Etiology

• Causal factor. Heterozygous germline BARD1 pathogenic variants, mostly truncating [memory].
• Breast cancer risk estimate [searched].
• Source class: primary literature, single cohort. Rofes et al., Genes 2021, DOI 10.3390/genes12020150 (PMC7911518), a Spanish hereditary breast and ovarian cancer cohort of 4,015 patients compared with gnomAD controls.
• Truncating variants overall: OR 3.78 (95% CI 2.10–6.48; p = 1.16 × 10⁻⁵).
• Hereditary breast cancer group: OR 4.18 (95% CI 2.10–7.70).
• Triple-negative breast cancer (TNBC) enrichment versus other subtypes: OR 5.40 (95% CI 1.77–18.15; p = 0.001).
• The authors conclude BARD1 is a moderate-penetrance breast cancer gene with a stronger TNBC association.
• A payer policy summary states BARD1 variants confer about 2–3× general-population breast cancer risk [searched: payer document, not primary]. This is lower than the single-cohort OR above. The two estimates differ in design and are not directly comparable.
• Large consortium estimates (Dorling and CARRIERS/BCAC-type studies, and the Weber-Lassalle series) were not retrieved. My search for them returned nothing relevant. Those are the primary sources the entry should cite for effect sizes.
• Pediatric arm: neuroblastoma [searched].
• Common-variant GWAS: Capasso et al. 2009, Nat Genet. Six SNPs at 2q35 in the BARD1 locus associated with high-risk neuroblastoma (397 cases/2,043 controls, replicated in 189 cases/1,178 controls). Source class: primary literature, common-variant association. It does not establish a high-penetrance Mendelian effect.
• Rare-variant enrichment: a cohort of 786 patients reported BARD1 as the most significantly enriched gene for pathogenic or likely pathogenic germline variants (OR 32.30; 95% CI 6.44–310.35), and linked germline P/LP variants to worse overall survival. Source class: primary literature, cohort. I saw only an abstract snippet, and the confidence interval is very wide.
• A bioRxiv preprint (2023) reports that common and rare neuroblastoma-associated BARD1 variants lower BARD1 mRNA and increase DNA damage, and that BARD1-LoF isogenic cells repair Cas9-induced damage less efficiently. Source class: preprint, not peer-reviewed in this retrieval. Check whether a peer-reviewed version exists.
• Environmental and protective factors: no BARD1-specific sources found. Do not import BRCA1/2 assumptions. Gene-environment interaction: no data retrieved.

3. Phenotypes

Everything below is [memory] unless stated. Frequencies and HPO bindings need checking against HPO, Orphanet and the primary literature.

• There is no known non-cancer syndromic phenotype. Do not invent a quality-of-life section; I have no data.
• Pediatric framing: onset age and penetrance for the neuroblastoma arm are not established in what I retrieved. Say so in the entry rather than imply a childhood surveillance indication.

4. Genetic and molecular information

• Gene: BARD1, chromosome 2q35. 2q35 is [searched] (Capasso GWAS). The HGNC ID (lowercase hgnc: form in this repo) must be looked up.
• Variants: mostly truncating or LoF. Missense variants in the RING and BRCT domains are largely VUS [memory]. Interpretation of missense variants is a known difficulty. I have no sourced classification data.
• Origin: germline. Mechanism: haploinsufficiency is supported for neuroblastoma-associated variants by the preprint [searched]. For breast cancer, LoF is the working model [memory].
• ClinGen validity: not found in my search, so no gene_disease_validity assertion can be copied. Check just genes-clingen-gaps and the cached CGGV records before writing one. Do not assign a tier.
• Modifier genes, epigenetics, chromosomal abnormalities: nothing retrieved.

5. Environmental information

None BARD1-specific retrieved. Leave environmental blocks empty rather than borrowing from BRCA1.

6. Mechanism (ordered causal chain)

1. A germline heterozygous LoF variant in BARD1 reduces BARD1 dosage. Haploinsufficiency is demonstrated in neuroblastoma-variant cell and expression data [searched, preprint]. It is inferred for breast cancer.
2. Reduced BARD1 destabilises or reduces the BRCA1–BARD1 heterodimer, which is [memory], mechanistically expected but not sourced here.
3. The result is a homologous recombination repair defect and increased DNA damage. The preprint supports reduced repair of Cas9-induced damage in BARD1-LoF cells [searched, in vitro].
4. Unrepaired damage and genomic instability follow. This is supported in mouse by McCarthy et al. 2003 (Mol Cell Biol; PMC162231 appeared in results): Bard1-null embryos die at E7.5–E8.5 from impaired proliferation, and mutant cells show increased aneuploidy [searched]. Evidence source: model organism.
5. Loss of the second allele or other cooperating events leads to tumour initiation. The two-hit step in human carriers is inferred, not shown in anything I retrieved.
6. The tissue pattern is breast, with a TNBC bias, and neuroblastoma in the pediatric arm. In mouse, mammary-specific conditional Bard1 loss produces basal-like mammary carcinomas with frequency, latency and histology indistinguishable from conditional Brca1 loss [searched; snippet only, the paper identity is "likely Shakya et al. 2008", unconfirmed]. This parallels the human TNBC enrichment. It is a model-organism result and the mouse is a full conditional knockout, not a heterozygous germline model.

Suggested GO/CL terms for this chain would be DNA repair and homologous recombination, ubiquitin ligase activity, and mammary epithelial cell and sympathetic neuroblast lineages. All are unverified and need ontology lookup.

7. Anatomical structures

Breast (mammary epithelium) and the sympathetic nervous system or adrenal medulla (neuroblastoma). Ovary is uncertain. No lateralisation data retrieved. UBERON/CL bindings: not looked up.

8. Temporal development

• Breast: adult onset. Specific age distribution not retrieved.
• Neuroblastoma: infancy to early childhood, as is general for the tumour [memory]. The age at diagnosis for BARD1 carriers specifically was not retrieved.

9. Inheritance and population

• Autosomal dominant with reduced penetrance [memory].
• Population frequency, founder variants, carrier frequency, sex ratio: not retrieved. gnomAD was used as the control set in the Rofes study [searched].
• Penetrance numbers: only the 2–3× relative-risk statement [searched, payer document] and the OR above. Absolute lifetime risk was not retrieved. The UK genomics hub text mentions a ≥30% estimated lifetime risk as a threshold for considering mastectomy [searched], which is a decision threshold and not a BARD1-specific penetrance figure.

10. Diagnostics

• Multigene hereditary cancer panel testing. BARD1 is included on common panels [memory].
• No BARD1-specific biomarker, imaging, or histology criteria retrieved, though TNBC histology is a reported association [searched].
• Variant interpretation: no ClinGen VCEP or ACMG-specific BARD1 guidance retrieved.

11. Prognosis

Not retrieved for adult breast carcinoma. For neuroblastoma, one cohort linked germline P/LP variants to worse overall survival [searched, abstract snippet].

12. Treatment and surveillance

• Surveillance: NCCN recommends annual mammogram and breast MRI from age 40 for carriers of BARD1, RAD51C and RAD51D [searched; quoted in an NSGC news item, secondary source. Verify against the NCCN guideline itself, which I did not open].
• Risk-reducing mastectomy: evidence described as insufficient or uncertain in several secondary sources (UK Genomics Education hub, FORCE, a payer summary) [searched]. The UK hub says it may be considered for higher-risk variants with an estimated lifetime risk of 30% or more.
• PARP inhibitors: BARD1 carriers' response is not established in anything I retrieved. I did not search for it, so do not assert either way.
• Candidate NCIT terms for surveillance and counselling procedures: not looked up. Use the repo's term-lookup workflow.
• Clinical trials: none searched.

13. Prevention

Enhanced breast screening as above. No evidence found for chemoprevention. Cascade testing and genetic counselling apply generally [memory]. For children, I found no surveillance recommendation for neuroblastoma in BARD1 carriers. The entry should record that gap.

14–15. Other species and model organisms

• Mouse: constitutive Bard1 knockout is embryonic lethal (E7.5–E8.5), delayed to about E9.5 when p53 is also removed [searched, McCarthy 2003, PMC162231]. Conditional mammary knockout gives basal-like carcinomas [searched, snippet; PMC2365565 appeared among results but I did not confirm it is that paper]. Conditional alleles exist (Columbia listing, MGI alleles MGI:2668394 and MGI:4939693 appeared in results, unverified).
• Fidelity caveat for a modeled_mechanisms link: the null and conditional-null models do not reproduce the human heterozygous moderate-penetrance state. They support the BRCA1–BARD1 tumour-suppressor mechanism but only partially recapitulate the human phenotype.
• Natural disease in other species, OMIA, zebrafish, Drosophila and cell lines: not searched.

Gaps to close before curation

1. Primary consortium effect sizes for breast cancer (BCAC/CARRIERS, Dorling, Weber-Lassalle). Not found in my search.
2. ClinGen gene-disease validity for BARD1 and its cached CGGV record.
3. Primary NCCN text, plus any ESMO/ACMG/ERN guidance.
4. Peer-reviewed status of the neuroblastoma preprint, and the exact citation for the 786-patient cohort.
5. Penetrance, ovarian and pancreatic risk, and tumour-spectrum data.
6. Identifiers: MONDO, OMIM, Orphanet, HGNC, and every HPO/GO/CL/UBERON/NCIT term.

Sources

• Rofes et al., Genes 2021, BARD1 and TNBC in a Spanish HBOC cohort (PMC7911518)
• Capasso et al. 2009 GWAS and later neuroblastoma results: bioRxiv preprint on BARD1 haploinsufficiency
• Capasso 2009 record, Birmingham
• Germline pathogenic variants in neuroblastoma enriched in BARD1 (Mount Sinai record)
• NSGC news item quoting NCCN screening
• UK NHS Genomics Education, BARD1
• FORCE, BARD1 risk management
• Payer policy 204126, germline genetic testing
• McCarthy et al. 2003, Bard1 knockout (PMC162231)
• Conditional Bard1 mammary tumour study (PMC2365565, unconfirmed)
• Columbia conditional Bard1 allele listing

Reference Validation

Checked with linkml-reference-validator 0.3.0.

All extracted references resolved successfully.

Term Validation

Checked with linkml-term-validator 0.4.5, through the ols: adapter.

Prefixes with no resolver

Terms carrying these prefixes were not checked either way, because no configured ontology covers them. An unrecognised prefix may name an ontology this run could not reach as easily as one that does not exist, so nothing here is evidence of fabrication: MGI.

1 of 3 terms resolved to a current term; the rest could not be looked up either way.