Mutational signatures are the fingerprints left behind by distinct biological processes. Ultraviolet light, for example, produces a characteristic pattern of cytosine-to-thymine changes in skin cancers, while defective DNA mismatch repair generates a different and highly recognizable signature in colorectal and other tumors. Each signature reflects a specific combination of insult and response: the chemical nature of the damage, the way the cell’s repair machinery processes that damage, and the errors that slip through when repair fails or is overwhelmed. Over the past decade, large international consortia have catalogued dozens of such signatures across cancer types, providing a reference framework against which individual tumors can be compared.

Prostate cancer presents a particularly interesting case for this kind of analysis. The disease is extraordinarily heterogeneous, ranging from indolent tumors that never require treatment to aggressive, metastatic cancers that resist hormonal therapy and become lethal. Genomic studies have shown that prostate tumors carry relatively few point mutations compared with some other cancers, but they are rich in structural rearrangements, copy-number changes, and complex events that shuffle large segments of the genome. Among these, the phenomenon of chromoplexy, a cascade of chained rearrangements, and the breakage-fusion-bridge cycles that generate extrachromosomal DNA have attracted particular attention as hallmarks of aggressive disease.

The central premise of the new study is that no single dimension of mutational data is sufficient to capture the full repertoire of processes operating in prostate tumors. Point mutations, small insertions and deletions, structural variants, and patterns of DNA methylation each carry partial information about the underlying biology. By integrating these data types into a unified analytical framework, the researchers aimed to define signatures that are more robust and more biologically informative than those derived from any one data type alone. This integrated approach reflects a broader trend in cancer genomics, where multi-modal signatures have proven capable of detecting processes that leave subtle or inconsistent traces in any single mutation class.

Technically, the construction of integrated signatures requires careful normalization of heterogeneous data. Single-base substitution signatures are typically represented as 96-channel vectors, capturing the trinucleotide context of each mutation. Indel signatures use a different set of categories based on sequence context and length, while structural variant signatures classify events by size, orientation, and the number of breakpoints involved. Copy-number signatures add yet another layer, encoding genome-wide patterns of segmental gains and losses. The analytical challenge lies in combining these disparate representations without allowing one data type to dominate the result, and in validating that the resulting signatures correspond to real biological processes rather than technical artifacts.

Applying this framework to prostate cancer cohorts, the study identifies a set of mutational processes that collectively define the disease. These include signatures associated with defective DNA repair, particularly involving homologous recombination, a pathway whose impairment in prostate cancer has major clinical implications because it sensitizes tumors to PARP inhibitors and platinum-based chemotherapy. Germline and somatic alterations in BRCA1, BRCA2, and ATM are well-established contributors to this phenotype, and signature-based detection offers a way to identify tumors with homologous recombination deficiency even when the causative alteration is not immediately apparent from sequencing data alone.

Beyond DNA repair deficiency, the integrated analysis illuminates processes tied to androgen receptor signaling and the genomic instability that accompanies disease progression. Androgen receptor activity is known to influence chromatin structure and replication timing, and there is growing evidence that it shapes the mutational landscape of prostate tumors in measurable ways. The study’s integrated signatures also capture age-related mutational processes, which accumulate slowly over decades and form the background against which disease-specific events unfold. Distinguishing this background from the processes that actively drive tumor evolution is essential for interpreting the genomic history of each cancer.

One of the most consequential aspects of signature analysis is its potential clinical utility. If a tumor’s signature profile can indicate which DNA repair pathways are compromised, it could help guide treatment decisions without requiring exhaustive functional testing. Signature analysis may also reveal exposure histories, such as prior chemotherapy or radiotherapy, whose mutational consequences persist in the tumor genome. In prostate cancer, where treatment sequencing remains a subject of intense debate, the ability to read a tumor’s mutational history could inform decisions about when to escalate therapy and which agents are most likely to be effective. The study’s authors position their integrated framework as a step toward this kind of clinically actionable interpretation.

The research also speaks to a persistent challenge in the field: reproducibility and standardization. Early studies of mutational signatures sometimes produced conflicting results, in part because different analytical pipelines could extract different numbers of signatures from the same data. Efforts by international consortia have since established reference catalogs and best practices, and the present study builds on this foundation by demonstrating that integrated, multi-modal signatures can be defined and validated in a disease-specific context. For prostate cancer, a tumor type whose genomic events are dominated by rearrangements rather than point mutations, this validation is particularly important, because methods optimized for other cancers may fail to capture the processes that matter most here.

Looking forward, the study suggests several avenues for further work. Larger and more diverse cohorts will be needed to confirm the findings across populations and disease stages, and longitudinal sampling of tumors before and after treatment could reveal how mutational processes evolve under therapeutic pressure. Combining signature analysis with single-cell and spatial genomic methods may add yet another dimension, allowing researchers to map mutational processes onto the architecture of tumors and their microenvironments. As the analytical tools mature, integrated mutational signatures are poised to become a standard component of prostate cancer genomics, bridging the gap between descriptive cataloguing and mechanistic understanding, and ultimately helping clinicians read each tumor’s history well enough to predict its future.

Subject of Research: Integrated multi-modal mutational signature analysis of prostate cancer genomes

Article Title: Integrated signatures define mutational processes in prostate cancer

Article References: Gruber, A. J., Olsen, A. V., Hernando, B., Cheng, K. C. L., Gerhäuser, C., Torres, M., Favero, F., Kiriy, D., Fernández-Sanromán, Á., Roldan-Romero, J. M., Barton, L., Pellegrina, D., Bova, G. S., Brewer, D. S., Brook, M. N., Brors, B., Butler, A., Cancel-Tassin, G., Corcoran, N. M., … Weischenfeldt, J. (2026). Integrated signatures define mutational processes in prostate cancer. Nature. https://doi.org/10.1038/s41586-026-10468-w

Image Credits: AI Generated

DOI: 10.1038/s41586-026-10468-w

Keywords: prostate cancer, mutational signatures, genomics, DNA repair, homologous recombination deficiency, structural variants, androgen receptor, chromoplexy, PARP inhibitors, tumor evolution, cancer genomics, Nature