How to Read BRCA1 and BRCA2 Results

A BRCA result is a statement about one or two specific DNA changes, classified on a five-tier scale. The tiers are pathogenic, likely pathogenic, uncertain significance (VUS), likely benign and benign. “Positive” means a pathogenic or likely pathogenic variant was found in BRCA1 or BRCA2. “Negative” means none was found in the regions the assay covered, which is not the same as none exists. A VUS means the lab found a change it cannot yet classify, and the correct handling of a VUS is to treat it as uninformative. Everything else on the report (risk percentages, surveillance language) is downstream interpretation of that one classification line. That interpretation belongs to a genetics clinician and does not belong to the assay.

The five tiers, and what they are built from

The classification comes from the ACMG/AMP framework, refined for BRCA1/2 by the ENIGMA consortium and ClinGen’s expert panel. Evidence codes get combined: null variants (nonsense, frameshift, canonical ±1/±2 splice) in a gene where loss of function is the disease mechanism carry strong weight, plus population frequency in gnomAD, segregation in families, multifactorial likelihood models, and functional assays such as saturation genome editing of BRCA1 exons.

Two practical consequences. First, tiers are reviewable: a variant classified as a VUS in 2015 may be likely benign now, and labs reissue reports. If you have an old result with a VUS, ask for a reclassification. Second, the classification is specific to the variant. It does not extend to the whole gene. “I have the BRCA gene” is not a meaningful statement. Everyone has two copies of BRCA1 and two of BRCA2, a gene first fully sequenced and annotated in the mid-1990s from a 117 kb genomic region on 17q.¹ What matters is which nucleotide changed.

The fields on the report that carry information

Read these, in order:

• Gene and transcript. BRCA1 variants should be described against NM_007294.4 (or ENST00000357654), BRCA2 against NM_000059.4. A c. coordinate without a transcript version is ambiguous.
• HGVS c. and p. notation, e.g. NM_007294.4:c.5266dupC (p.Gln1756Profs*74). Legacy names (185delAG, 5382insC, 6174delT) still appear on older reports. These map to BRCA1:c.68_69delAG, BRCA1:c.5266dupC, and BRCA2:c.5946delT, the three Ashkenazi founder variants.
• Classification, plus the date and the classifying source.
• Methods and limitations. This is the paragraph most people skip and the one that determines what “negative” means. Look for: sequencing of coding exons plus ±20 bp of flanking intron, and whether large rearrangement analysis (MLPA or an equivalent) was performed. Diagnostic NGS workflows for these two genes are explicitly designed with a separate dosage arm for exon-level copy number, because sequence calling alone does not see it.²
• Zygosity and allele fraction. Germline heterozygous calls should sit near 50% variant reads. A variant at 8% read support in blood is more likely clonal hematopoiesis or an artifact than a germline finding.

If the report came from a consumer array, the limitation section is the whole story. Array-based BRCA reporting typically genotypes three specific Ashkenazi founder variants. Thousands of pathogenic BRCA1/2 variants exist, and a negative array result excludes almost none of them.

Pulling the same calls out of your own data

If you have a WGS BAM/CRAM and a VCF, you can reproduce the sequence-level part of the analysis. GRCh38 coordinates:

BRCA1  chr17:43,044,295-43,125,364   (minus strand)
BRCA2  chr13:32,315,086-32,400,268   (plus strand)

Coverage first, because interpretation of a negative depends on it:

mosdepth --by brca_exons.bed --thresholds 1,10,20,30 brca sample.cram
# then check: any exon with <20x mean, or <95% of bases at >=20x, is not callable

Then annotate. We use VEP with an explicit transcript pick rather than letting it report every isoform:

vep -i sample.vcf.gz -o brca.vep.vcf --vcf --cache --assembly GRCh38 \
    --transcript_version --hgvsg --hgvs --symbol --canonical \
    --plugin SpliceAI,snv=spliceai_snv.hg38.vcf.gz,indel=spliceai_indel.hg38.vcf.gz \
    --custom clinvar.vcf.gz,ClinVar,vcf,exact,0,CLNSIG,CLNREVSTAT \
    --pick_allele_gene

Then filter to the two genes and read every coding or splice-region variant by hand. There will be on the order of tens of them across both genes, most common and benign. The rule we use: a variant is worth attention if gnomAD popmax allele frequency is below about 0.1% and either ClinVar carries a pathogenic or likely pathogenic assertion with two-star-or-better review status, or the consequence is a frameshift/nonsense/canonical splice change, or SpliceAI delta score exceeds ~0.5. CLNREVSTAT matters as much as CLNSIG: a single-submitter “pathogenic” with no assertion criteria is weak evidence.

Do not stop at SNVs and small indels. Run a structural variant caller and a read-depth CNV caller over the same interval, because Alu-mediated exon-level deletions and duplications in BRCA1 are a real fraction of pathogenic findings in some populations. Manta plus GATK gCNV, or CNVkit on targeted data, will find most multi-exon events. Panel-based capture workflows sized to a few hundred kb can recover these when paired with dosage analysis.³

What short reads miss, and where the analysis fails

• Large rearrangements. This is the biggest gap. Breakpoints inside repetitive Alu sequence are hard to resolve with 150 bp reads, and a clean sequence VCF says nothing about a deleted exon 13. Clinical labs still use MLPA for this reason.⁴
• Homopolymers and short tandem repeats. Indel calls in BRCA2 exon 11 (a very large exon with low-complexity stretches) are the most common source of false positives. Check them in IGV before you believe them.
• Deep intronic and regulatory variants. Standard reporting covers exons plus ~20 bp. Deep intronic changes that create new splice acceptors are usually not reported, though RNA evidence can resolve them.
• Mosaicism and low-fraction variants, which need higher depth than typical 30x WGS.

Sensitivity and specificity for BRCA NGS are workflow properties, not sequencer properties: validated diagnostic pipelines are built and benchmarked against Sanger and dosage reference sets before use.²⁵ Founder-variant work in specific populations shows how much the pre-test allele spectrum shapes what an assay should be designed to detect.⁶ Research-grade output from your own data is a hypothesis. Anything you would act on has to be confirmed in a CLIA/CAP-accredited lab on a fresh sample.

Penetrance: what a positive result predicts

For BRCA1 pathogenic variant carriers, cumulative breast cancer risk to age 80 is roughly 60-70%, and ovarian cancer risk roughly 40%. For BRCA2, breast cancer is similar to slightly lower and ovarian is substantially lower, on the order of 15-20%. BRCA2 also carries elevated male breast, prostate, and pancreatic cancer risk. Those are population averages. Individual risk is modified by variant position, family history, polygenic background, and age, which is why a number on a report is not a prediction for you.

Tumor biology differs too. BRCA1-associated tumors skew triple-negative and show characteristic homologous-recombination-deficiency mutational signatures on whole-genome sequencing of the tumor.⁷ Genome profiling of premalignant tissue from carriers is an active research direction for understanding when those changes appear.⁸ Note the distinction: germline testing asks about the DNA you inherited, tumor sequencing asks what a specific tumor has done.

Questions people also ask

Do I need a mastectomy if I have a BRCA variant? No. A pathogenic variant is a risk estimate. The management options are enhanced surveillance with MRI, risk-reducing surgery and risk-reducing medication. Choosing among them is a decision for you and a genetics and oncology team, weighing your age, variant, family history, and preferences. Evidence on how BRCA results change surgical decisions in newly diagnosed patients is mixed, with a measurable but not universal effect.⁹ Genomic information has been reshaping these conversations for well over a decade, and the constraint has always been interpretation rather than sequencing.¹⁰

Which is more serious, BRCA1 or BRCA2? BRCA1 carries higher average ovarian cancer risk and earlier-onset, more often triple-negative breast cancer. BRCA2 carries a broader spread of other cancers. Neither label substitutes for the specific variant and your family history.

Which parent carries it, mom or dad? Either. BRCA1 and BRCA2 are autosomal, so a variant is inherited from one parent regardless of sex, and fathers transmit it as often as mothers. A paternal BRCA2 variant is easy to miss in a family history that only looks at women.

Can one sibling have it and the other not? Yes. Each child of a carrier has a 50% chance of inheriting the variant, independently. Siblings with different results is the expected outcome. It is not an error.

Is it better to be BRCA negative or positive? A true negative is better in terms of risk. Read the limitations section before trusting a negative: an array that tested three variants, or a sequencing report without large rearrangement analysis, leaves most of the gene unexamined. And a negative BRCA result leaves other genes unaddressed, which is why panels rather than two-gene tests are now standard. It does not address the following genes.³

• PALB2
• ATM
• CHEK2
• TP53
• RAD51C
• RAD51D

What should I do first with a positive result? Get a genetic counselor. They confirm the classification against current expert-panel assertions, work out who in your family should be tested for that specific variant, and map surveillance options. This is the one step we would not skip.

Footnotes

1. T M Smith, M K Lee, C I Szabo, et al. Complete genomic sequence and analysis of 117 kb of human DNA containing the gene BRCA1. Genome Research, 1996. https://doi.org/10.1101/gr.6.11.1029 ↩

2. Lídia Feliubadaló, Adriana Lopez-Doriga, Ester Castellsagué, et al. Next-generation sequencing meets genetic diagnostics: development of a comprehensive workflow for the analysis of BRCA1 and BRCA2 genes. European Journal of Human Genetics, 2012. https://doi.org/10.1038/ejhg.2012.270 ↩ ↩²

3. Laurent Castéra, Sophie Krieger, Antoine Rousselin, et al. Next-generation sequencing for the diagnosis of hereditary breast and ovarian cancer using genomic capture targeting multiple candidate genes. European Journal of Human Genetics, 2014. https://doi.org/10.1038/ejhg.2014.16 ↩ ↩²

4. Andrew J Wallace. New challenges for BRCA testing: a view from the diagnostic laboratory. European Journal of Human Genetics, 2016. https://doi.org/10.1038/ejhg.2016.94 ↩

5. S M Pilgrim, S J Pain, M D Tischkowitz. Opportunities and challenges of next-generation DNA sequencing for breast units. British Journal of Surgery, 2014. https://doi.org/10.1002/bjs.9458 ↩

6. Ava Kwong, Enders Kai On Ng, Chris Lei Po Wong, et al. Identification of BRCA1/2 Founder Mutations in Southern Chinese Breast Cancer Patients Using Gene Sequencing and High Resolution DNA Melting Analysis. PLoS ONE, 2012. https://doi.org/10.1371/journal.pone.0043994 ↩

7. Johan Staaf, Dominik Glodzik, Ana Bosch, et al. Whole-genome sequencing of triple-negative breast cancers in a population-based clinical study. Nature Medicine, 2019. https://doi.org/10.1038/s41591-019-0582-4 ↩

8. Takeshi Hirose, Masachika Ikegami, Kumiko Kida, et al. Cancer risk assessment of premalignant breast tissues from patients with BRCA mutations by genome profiling. npj Breast Cancer, 2024. https://doi.org/10.1038/s41523-024-00693-9 ↩

9. Quratul Ain, Caroline Richardson, Miriam Mutebi, et al. Does mainstream BRCA testing affect surgical decision-making in newly-diagnosed breast cancer patients?. The Breast, 2023. https://doi.org/10.1016/j.breast.2022.12.001 ↩

10. Dimosthenis Ziogas, Dimitrios H. Roukos. Genetics and Personal Genomics for Personalized Breast Cancer Surgery: Progress and Challenges in Research and Clinical Practice. Annals of Surgical Oncology, 2009. https://doi.org/10.1245/s10434-009-0436-2 ↩