DNA is continuously exposed to chemical damage. Thousands of reactions occur inside every cell each day, and some of them alter DNA bases, break chemical bonds or interfere with the normal pairing of the two strands. Cells possess several repair systems that patrol the genome and correct many of these lesions. But repair is not perfect. Some damaged sites are recognized quickly and repaired efficiently, while others remain undetected long enough to be copied during cell division. Once a damaged base is converted into a different sequence through replication or faulty repair, the change may become permanent. “The rate at which mutations accumulate is a balance between the rate of damage and the rate of repair,” explains Dr. Ariel Afek, whose laboratory led the study. That balance is not uniform across the genome, and the new work helps reveal why.

Most research on genome instability has focused on mutations that are easy to observe because they remain as lasting changes in DNA. Afek’s team instead examined the temporary lesions that precede those changes. This distinction is important because a damaged DNA base does not inevitably become a mutation. Its fate depends on whether repair enzymes locate it, bind to it and remove it before the cell copies the damaged strand. To investigate these early steps, the researchers created a molecular chip containing thousands of short DNA molecules. Each molecule carried the same type of artificial damage, but the surrounding DNA letters were varied. This design allowed the scientists to compare repair activity at many sequence contexts while keeping the central lesion constant.

The experiments showed that the repair enzymes did not treat every damaged site equally. Their ability to recognize and bind the lesion depended on the precise combination of bases around it. The influence extended as far as five DNA positions upstream or downstream from the damaged base, indicating that the enzymes read a much larger molecular environment than the lesion alone. Noga Levy, a doctoral student in Afek’s laboratory and the study’s lead researcher, compared this behavior to an editor evaluating a word in context rather than in isolation. A damaged base may be chemically identical in two locations, yet the surrounding sequence can determine whether a repair protein notices it efficiently or passes it by.

The sequence effect was not simply a matter of the letters themselves. The researchers found that preferred DNA sequences shared physical characteristics, including distinctive shapes in the double helix. DNA is often represented as a uniform spiral staircase, but its structure changes subtly from one sequence to another. Some combinations of bases bend more easily, widen or narrow the grooves on the helix, or alter the distribution of electrical charge along the molecule. One of the repair enzymes studied by the team favored damaged sites embedded in sequences that create an unusually narrow region of the double helix. Such structural variation can provide a recognition signal that complements the chemical features of the damaged base.

To understand the molecular basis of this preference, the Weizmann researchers collaborated with a group led by Prof. Brian P. Weiser at Rowan University in New Jersey. Using computer simulations, the scientists examined how the repair enzyme moved across DNA and interacted with the region surrounding the lesion. The simulations indicated that one amino acid in the enzyme scans the local DNA structure. It is attracted to the negative electrical charge associated with the narrow helical region, helping guide the protein toward sequences with the appropriate shape. This model illustrates how DNA repair can depend on both chemistry and mechanics: the enzyme is not only searching for a damaged base, but also sensing the architecture and electrostatic landscape of the surrounding double helix.

The team then asked whether these biochemical preferences could be detected in the human genome after millions of years of evolution. One common form of genomic damage occurs when a cytosine, or C base, is chemically altered and no longer pairs correctly with guanine. Several important repair enzymes identify and remove the incorrect base, restoring the proper sequence. The researchers reasoned that genomic regions where a particular repair enzyme operates efficiently should preserve more cytosines, because damage at those sites would be more likely to be corrected. In regions where repair is inefficient, comparable damage should more often escape correction and accumulate as mutations. Analysis of genomic data revealed a correlation consistent with this prediction for one of the enzymes. The result suggests that repair preferences are not merely laboratory curiosities; over evolutionary time, they can influence which sequences remain stable and which become progressively altered.

That observation has implications for interpreting human evolution. When scientists identify genetic changes that became common in modern humans, they often ask whether those changes were favored because they helped people adapt to environmental pressures. But not every widespread change necessarily reflects natural selection. Some may have accumulated because the DNA sequence was especially vulnerable to damage or because repair enzymes were less effective in that context. “To identify which genomic changes were adopted by humans to survive a changing environment, we must first understand which changes accumulate naturally due to the preferences of the repair mechanisms,” Afek says. Distinguishing selection from repair-driven mutation could make evolutionary analyses more precise, particularly in genomic regions with unusually high or low mutation rates.

The same principle may help explain the genetic history of cancer. A tumor typically begins when one cell accumulates mutations that alter growth control, DNA maintenance or communication with neighboring cells. As that cell divides, additional changes form characteristic combinations known as mutational signatures. These signatures can reveal the kinds of damage that occurred and the repair pathways that were active or defective. In the new study, the researchers found a relationship between the sequence preferences of the repair enzymes and mutation patterns observed in human tumors. One explanation is that damage to repair systems in cancer cells allows mutations to accumulate in genomic regions that were previously protected. Another is that evolution has tuned repair enzymes to recognize regions that are intrinsically more vulnerable. Although the study does not establish a single cause, it reinforces the idea that repair activity is a major force shaping cancer genomes.

The findings may ultimately have practical applications beyond understanding mutation. DNA repair enzymes are already used in biotechnology and gene-editing systems, where their ability to recognize particular structures can be harnessed to modify genetic material. Mapping the sequence and shape preferences of these proteins could allow researchers to design more precise molecular tools or engineer enzymes that protect vulnerable regions of the genome more effectively. Afek’s laboratory is extending the approach to additional repair mechanisms in work led by graduate student Noga Carmon. By studying repair as a process that combines damage recognition, sequence context, molecular shape and electrical charge, scientists may gain a more complete view of how cells preserve genetic information—and how failures in that preservation can drive disease.

Web References: Nature Communications article; DOI link

References: Nature Communications, DOI: 10.1038/s41467-026-74090-0

Keywords: DNA repair, mutations, genome stability, genetic damage, structural biology, DNA sequence context, cancer genomics, mutational signatures, evolution, gene editing

CUP is notoriously challenging in clinical practice due to its ambiguous nature; metastatic tumors are identified, yet doctors remain unable to locate the original tumor site. This uncertainty hampers precise treatment planning, leading to empiric chemotherapy regimens that often fall short of improving prognosis. The researchers tackled this ongoing dilemma by employing WGS as a diagnostic tool, leveraging its unparalleled ability to decode the complete genetic blueprint of tumor cells and uncover hidden mutational patterns characteristic of specific tissues.

By analyzing genome-wide mutations, structural variants, and mutational signatures through sophisticated computational frameworks, the team successfully linked metastatic samples to their most probable tissue of origin. The granularity afforded by WGS data, including single nucleotide variants, copy number alterations, and chromosome rearrangements, provides a genomic fingerprint uniquely reflective of the tumor’s biological genesis. This methodological advancement represents a departure from traditional immunohistochemistry and targeted gene panels, which provide limited molecular insights and often fail to resolve cases.

The study encompassed a large cohort of CUP patients whose clinical diagnostics had yielded inconclusive or ambiguous results. Each tumor specimen underwent high-depth WGS, followed by bioinformatic integration with extensive reference databases containing genomic profiles from various known cancer types. Advanced machine learning algorithms parsed through these complex datasets, identifying patterns consistent with specific cancer lineages, even in highly heterogeneous and evolutionarily dynamic metastatic tissues.

Notably, the improved accuracy in diagnosing tissue of origin translated into tangible clinical benefits. Armed with genomic evidence indicative of the primary site, oncologists could tailor treatment regimens more precisely, aligning therapies with those used for anatomically defined cancers. This personalization has direct implications for patient outcomes, as tissue-specific treatment protocols often outperform empirical chemotherapy applied to CUP cases treated blindly.

The researchers emphasize that WGS provides a panoramic view of tumor biology that surpasses the capabilities of conventional diagnostic modalities. For instance, mutational signature analyses reveal underlying carcinogenic processes, such as tobacco exposure or UV damage, which serve as indirect indicators of tumor provenance. Structural variations and chromosomal abnormalities further refine classification, enabling the distinction between morphologically similar but molecularly distinct cancers.

Importantly, the study also found that some tumors classified as CUP harbored actionable mutations, opening new avenues for targeted therapies. The integration of WGS in diagnostic pipelines allows clinicians to not only identify the cancer’s origin but also detect genomic alterations susceptible to existing molecularly targeted agents or immunotherapies. This dual utility enhances the potential to administer precision oncology, moving beyond simply guessing the tumor type toward actively exploiting its vulnerabilities.

From a technical perspective, the implementation of WGS posed challenges such as data complexity, interpretation hurdles, and the need for rapid turnaround times compatible with clinical workflows. The research team addressed these obstacles by optimizing sequencing protocols, employing streamlined bioinformatic tools, and validating their findings across multiple independent cohorts to ensure reproducibility and robustness.

Moreover, the study paves the way for integrating WGS into standard-of-care practices, highlighting how genomics-driven diagnostics could become fundamental in managing CUP and perhaps other diagnostically challenging cancers. The cost-effectiveness of WGS continues to improve as sequencing technologies advance and computational infrastructures become more accessible, reinforcing its feasibility for widespread clinical use.

The implications extend beyond diagnostic clarity; uncovering the primary tumor source allows for better prognostic assessments. Prognostic biomarkers derived from WGS data can stratify patients based on expected disease trajectories, enabling more informed decisions about treatment intensity and follow-up strategies. Accurately identifying the tissue of origin also facilitates enrollment in clinical trials targeting specific cancer types, broadening therapeutic options for CUP patients.

Intriguingly, the study’s findings resonate with emerging paradigms in oncology that recognize cancer as a genomic disease defined by its mutation landscape rather than solely by histopathology. The ability of WGS to capture the full spectrum of genomic alterations empowers oncologists to transcend traditional classification systems steeped in morphology and immunophenotyping, moving toward molecular taxonomy that better reflects tumor biology.

This research underscores the transformative potential of genomics-driven precision medicine in oncology, especially for complex cases like CUP where uncertainty has long hindered progress. By unlocking the molecular secrets encoded in tumor genomes, WGS embodies a new frontier in cancer diagnosis and treatment, promising to shift clinical paradigms and improve survival outcomes through tailored therapeutic approaches.

As genomic databases continue to expand and machine learning algorithms become more sophisticated, the accuracy and utility of WGS in cancer diagnostics will only grow. The study by Rebello and colleagues stands as a compelling proof-of-concept that integrating whole genome sequencing into clinical practice is not only feasible but highly beneficial, setting a roadmap for future innovations in cancer care.

Ultimately, the integration of comprehensive genomic technologies heralds a more hopeful era for CUP patients, who have historically faced grim prognoses due to diagnostic ambiguity. WGS offers a beacon that illuminates the hidden origins of metastatic cancers, enabling clinicians to deploy more effective, targeted interventions based on precise molecular diagnoses rather than trial-and-error approaches.

Through this innovative research, the vision of personalized oncology comes closer to reality, where every cancer patient’s treatment plan is informed by the unique genetic architecture of their tumor. As the field continues to evolve, whole genome sequencing is poised to become an indispensable tool in the oncologist’s arsenal, catalyzing a new epoch of precision diagnostics and individualized therapeutics that improve lives and redefine cancer care.

Subject of Research: Whole genome sequencing application in tissue-of-origin diagnosis and treatment optimization for cancer of unknown primary (CUP).

Article Title: Whole genome sequencing improves tissue-of-origin diagnosis and treatment options for cancer of unknown primary.

Article References:
Rebello, R.J., Posner, A., Dong, R. et al. Whole genome sequencing improves tissue-of-origin diagnosis and treatment options for cancer of unknown primary. Nat Commun 16, 4422 (2025). https://doi.org/10.1038/s41467-025-59661-x

Image Credits: AI Generated