The scientific backdrop to this study is a genuine epidemiological puzzle. For decades, researchers have known that women who have given birth to more children experience a lower incidence of breast cancer after menopause, with the risk declining roughly linearly as parity increases. Pregnancy induces profound and lasting changes in breast tissue, including differentiation of the mammary epithelium, and many investigators have assumed that these changes would be reflected in altered gene expression profiles that could be detected decades later. Yet the precise molecular mechanism by which childbearing confers its protective effect has remained elusive, and attempts to pin it down at the level of tissue transcriptomics have produced inconsistent results.

The NOWAC research group, led by Eiliv Lund of UiT The Arctic University of Norway in Tromsø, had previously reported a striking observation that seemed to point toward a mechanism. In studies of peripheral blood cells from healthy postmenopausal women, the team demonstrated that increasing parity was associated with a linear decrease in the expression of hundreds of genes. This gene expression signature in blood cells appeared to mirror the linear decline in postmenopausal breast cancer incidence with increasing parity. Intriguingly, however, the same parity-related expression changes were not observed in the blood cells of women who had already developed breast cancer, suggesting that the signature might represent a protective state that is lost or overridden during carcinogenesis.

Those earlier findings raised an obvious and important question: if parity reshapes gene expression in blood cells in a way that tracks cancer risk, does it also reshape gene expression in the breast tissue itself, where tumors actually arise? Answering this question required tissue samples that are extraordinarily difficult to obtain from healthy women, because there is ordinarily no clinical reason to biopsy normal breast tissue. The NOWAC infrastructure, a prospective cohort study that enrolled 172,000 women between 1991 and 2007, provided the framework needed to attempt it.

Between 2006 and 2010, the researchers recruited women in two ways. Women enrolled in NOWAC who received a diagnostic biopsy for breast cancer were asked to provide a second biopsy specifically for research purposes, with the collaboration of surgeons at ten hospitals participating in the Norwegian Breast Cancer Group. In parallel, normal breast tissue was collected from NOWAC participants undergoing the national mammographic screening program, an effort made possible by a feasibility study showing that 74 percent of healthy women approached at a screening center in Tromsø agreed to donate tissue. The final analytical dataset consisted of 279 age-matched case-control pairs, and the paired matched design was maintained throughout all laboratory analyses, a methodological choice that strengthens the comparison between cancer patients and healthy controls.

The laboratory work relied on Illumina microarray technology to measure genome-wide gene expression, with differential expression analysis performed using the limma Bioconductor package in R, a standard and widely trusted tool for microarray statistics. To characterize the molecular subtype of each tumor and each normal tissue sample, the researchers applied the PAM50 test, a 50-gene classifier that assigns breast tumors to intrinsic subtypes such as luminal A, luminal B, HER2-enriched, and basal-like. The statistical framework controlled the false discovery rate at a q-value threshold of 0.05, ensuring that reported findings were unlikely to be artifacts of multiple testing across thousands of genes.

The results on the cancer side were dramatic, but not in the way the parity hypothesis would predict. When the researchers compared breast cancer tissue with normal breast tissue, they found that nearly the entire transcriptome was differentially expressed: 10,013 of the 11,308 genes assessed showed statistically significant differences between cases and controls. Of these, 5,768 genes were upregulated and 4,245 genes were downregulated in cancer tissue, a sweeping molecular signature of the malignant state that confirms the biological gulf between tumor and normal tissue. This massive case-control signal also demonstrates that the study had ample statistical power to detect expression differences, which makes the central null finding all the more meaningful.

That central finding is the study’s headline result. When parity, classified into three categories of zero children, one to three children, and four to eight children, was incorporated into the statistical model as an explanatory variable, only three genes in breast cancer tissue showed significant changes associated with parity, and not a single gene in normal breast tissue did. Gene set enrichment analyses, which test whether predefined biological pathways are collectively altered, returned non-informative results. Furthermore, the PAM50 classification failed to establish any relationship between parity and the distribution of intrinsic molecular subtypes in either tissue type. In other words, across the entire measurable transcriptome, the number of full-term pregnancies left virtually no trace in the breast.

The authors conclude that their study revealed no significant differences in gene expression between normal and breast cancer tissue in relation to the number of full-term pregnancies. This is a genuine null result from a well-powered, carefully matched design, and null results of this kind carry real scientific weight. They effectively rule out a whole class of mechanistic explanations: whatever protects parous women from postmenopausal breast cancer, it does not appear to operate through durable, detectable changes in the steady-state expression levels of genes in breast tissue, whether healthy or malignant. The protective signal that the team previously observed in peripheral blood cells now stands in sharper contrast, suggesting that the parity-related biology may be systemic rather than resident in the breast itself, or that it operates through mechanisms invisible to bulk transcriptomics, such as epigenetic marks, stromal architecture, immune surveillance, or transient changes that fade long before menopause.

The study also stands as a testament to the value of long-term prospective cohorts with biobanks. Only a resource like NOWAC, which links decades of lifestyle and reproductive history data to consented tissue donations and national cancer registry records, could assemble 279 matched pairs of cancer and normal breast tissue with complete parity information. The work was supported by the European Research Council, the Northern Norway Regional Health Authority, and the Norwegian Cancer Society, and the authors emphasize that the funders played no role in the design, analysis, or interpretation of the study. For researchers hunting the molecular basis of pregnancy’s protective effect, the message is clear: the answer will not be found by simply cataloging which genes are switched on or off in breast tissue. The search must now move to other layers of biology, and this study, by closing one door definitively, helps point the way.

Subject of Research: The effect of parity on global gene expression in normal breast tissue and breast cancer tissue

Article Title: No impact of parity on global gene expression levels in breast cancer tissue and normal breast tissue – a nested case-control study in the Norwegian Women and Cancer study

Article References: Lund, E., Krum-Hansen, S., Olsen, K. S., Shvetsov, N., Snapkow, I., Gavriluk, O., Busund, L.-T. R., Frantzen, J. O., Holden, M., & Holden, L. (2026). No impact of parity on global gene expression levels in breast cancer tissue and normal breast tissue – a nested case-control study in the Norwegian Women and Cancer study. BMC Cancer. https://doi.org/10.1186/s12885-026-17044-5

Image Credits: AI Generated

DOI: 10.1186/s12885-026-17044-5

Keywords: breast cancer, parity, gene expression, NOWAC study, transcriptomics, microarray, PAM50, nested case-control, epidemiology, reproductive history, cancer prevention, Norway

The framework, called HCAM-BRCA, short for Hierarchical Cross-Attention Multi-omics, tackles a problem that has dogged computational oncology for years: how to combine different kinds of molecular data without losing the biological relationships between them. Modern cancer research generates information from multiple molecular layers simultaneously. Messenger RNA profiles reveal which genes are being actively transcribed into protein-building instructions. MicroRNA profiles capture short regulatory molecules that silence gene expression after transcription. DNA methylation data, measured at cytosine-phosphate-guanine sites across the genome, show chemical tags that can switch genes on or off without altering the underlying sequence. Each layer tells part of the story, and each layer regulates the others in a web of interactions that single-omics analyses simply cannot see.

Many existing integration approaches handle this complexity in limited ways. Some methods analyze one data type at a time, while others combine only two layers in pairwise fashion, for instance joining RNA and microRNA data or RNA and methylation data but never all three at once. More importantly, most approaches do not explicitly model the hierarchical regulatory structure that connects the layers: methylation influences transcription, microRNAs modulate messenger RNA, and feedback loops run in both directions. HCAM-BRCA was designed to capture exactly these relationships. The architecture applies modality-specific self-attention to each data type, allowing the model to learn which features within a layer depend on one another, and then deploys cross-attention mechanisms between layers so the model can learn how, for example, methylation patterns inform the interpretation of gene expression.

Attention mechanisms are the same mathematical machinery that powers modern large language models, and their great advantage here is interpretability. Rather than acting as an inscrutable black box, an attention-based model assigns weights that reveal which features it considered most important when making a decision. In HCAM-BRCA, those weights translate directly into a ranked list of genes and regulatory elements that drove each subtype classification. The learned low-dimensional embeddings, compact numerical representations of each tumor’s multi-omics profile, were then fed into a battery of conventional machine learning classifiers, allowing the researchers to test how well the attention-derived representation supported downstream prediction.

The results were evaluated on data from The Cancer Genome Atlas, the large public repository of de-identified tumor profiles that has become the workhorse of computational cancer research. Performance was measured with rigorous statistics, including the Matthews correlation coefficient, a demanding metric that accounts for class imbalance, and the area under the precision-recall curve, which is particularly informative when one subtype is rare. All metrics were reported as means with standard deviations across five-fold cross-validation, meaning the data were repeatedly split so that every model was tested on samples it had never seen during training. The researchers also addressed a common pitfall in cancer genomics: class imbalance, where some subtypes have far fewer samples than others. Using an adaptive synthetic sampling technique known as ADASYN, they generated balanced training sets to prevent the models from simply defaulting to the most common categories.

Across every classifier tested, HCAM-BRCA’s three-layer integration outperformed the pairwise RNA-microRNA and RNA-CpG strategies, confirming that the full hierarchical picture carries information that two-layer views miss. Among the twelve machine learning models evaluated, CatBoost, a gradient-boosting method, achieved the highest macro-average Matthews correlation coefficient at 0.7628 with a standard deviation of 0.0130, while Extra Trees attained the highest area under the precision-recall curve at 0.8925 with a standard deviation of 0.0032. Those numbers represent strong and stable discrimination across subtypes, and notably the framework performed well precisely where existing tools struggle most, in separating the clinically challenging Luminal B and Her2-enriched categories.

Accuracy alone, however, was only half the story. Because the model’s attention weights expose which features influenced each decision, the researchers could interrogate the biology behind the predictions. Attention-based feature prioritization surfaced genes associated with transcriptional regulation, chromatin organization, immune signaling, and established cancer-related pathways, a pattern consistent with what independent functional analyses using Gene Ontology and KEGG pathway annotations would expect from genuinely relevant candidates. In other words, the model was not latching onto statistical noise; it was converging on genes that biologists already recognize as players in tumor behavior, along with others that merit new investigation.

The most clinically provocative findings came from survival analysis. The team examined whether expression of the prioritized genes was associated with patient outcomes within specific subtypes, and several strong subtype-specific prognostic links emerged. In Basal-like tumors, the aggressive category that largely overlaps with triple-negative breast cancer, the genes CUX1, MYZAP, and CEBPA showed significant survival associations. In Luminal B tumors, HOXA10, a developmental regulator frequently implicated in cancer, and HLA-DQA2, an immune-related gene, carried prognostic weight. In Her2-enriched tumors, GDF10 and ZNF879 were associated with survival outcomes. Each of these associations suggests a potential biomarker that could, with further validation, help stratify patients within a subtype for more tailored follow-up or therapy.

The significance of such subtype-specific markers is easy to underestimate. A gene that predicts poor survival in Basal-like tumors may be irrelevant in Luminal B, and pooling all subtypes together in a single analysis can wash out these signals entirely. By classifying first and then probing survival within each molecular category, HCAM-BRCA mirrors the way precision oncology actually operates: treatment decisions are made for a specific patient with a specific tumor subtype, not for an average cancer. A framework that simultaneously classifies accurately and nominates candidate biomarkers within each class therefore addresses two bottlenecks in the same pipeline.

The study, published open access on 23 September 2026 with no external funding, arrives amid a broader wave of multi-omics integration methods, including graph convolutional approaches such as MOGONET and neural frameworks like moBRCA-net. What distinguishes HCAM-BRCA is its explicit attention to hierarchical cross-layer regulation and the interpretability that attention weights provide. The authors acknowledge that the work rests on publicly available TCGA data and used no new human subjects, which means the biomarker candidates remain hypotheses awaiting validation in independent cohorts and, eventually, clinical settings. Still, the combination of robust subtype classification, biologically coherent feature prioritization, and subtype-specific prognostic signals makes a compelling case that teaching artificial intelligence to read all three molecular layers of a tumor at once, and to explain what it read, could move computational oncology closer to the clinic. For patients facing the diagnostic gray zones that complicate breast cancer care today, tools that sharpen those boundaries while revealing the genes behind them are exactly the kind of advance the field has been waiting for.

Subject of Research: Interpretable multi-omics integration using hierarchical cross-attention for breast cancer subtype classification and biomarker discovery

Article Title: HCAM-BRCA: an interpretable multi-omics framework for breast cancer subtype classification and biomarker discovery

Article References: Ali, M. N., & Biswas, S. (2026). HCAM-BRCA: an interpretable multi-omics framework for breast cancer subtype classification and biomarker discovery. BMC Bioinformatics. https://doi.org/10.1186/s12859-026-06607-9

Image Credits: AI Generated

DOI: 10.1186/s12859-026-06607-9

Keywords: breast cancer, multi-omics integration, cross-attention, machine learning, biomarker discovery, TCGA, subtype classification, PAM50, prognostic genes, precision oncology, DNA methylation, BMC Bioinformatics