Study Reveals Hidden Mutational Process in Colorectal Cancer

By Dr. Chinta SidharthanReviewed by Laura ThomsonSep 3 2026

Roughly 10% of colorectal tumors without obvious defects in their deoxyribonucleic acid (DNA) repair machinery still carry mutations that resemble the fingerprints of repair failure. In a recent study published in Nature Communications, scientists analyzed thousands of whole-genome sequences to determine whether one of these previously overlooked mutation patterns actually represents a distinct process in colorectal cancer (CRC).

Introduction

Mutational signature analysis has become a routine tool in cancer genomics, allowing researchers to trace the biological and environmental processes that leave characteristic patterns of DNA damage across a tumor's genome. Large reference catalogs and databases compile known signatures linked to various contributing processes, including aging and DNA repair defects, as well as various environmental exposures.

When researchers detect a new pattern in a tumor cohort, they typically try to explain it as a combination of these established signatures, a process called decomposition, rather than assuming it reflects an entirely novel mutational process. This conservative approach helps avoid overstating discoveries.

However, it also risks dismissing patterns as mixtures of familiar signatures even when the evidence is ambiguous. Distinguishing overlapping signatures remains especially difficult in CRC, where patterns resembling DNA repair deficiency can be easy to misread.

Want to save this article for later? Click here.

The Current Study

In the present study, the researchers revisited whole-genome sequencing data from 802 treatment-naive, DNA repair-proficient, microsatellite-stable colorectal tumors previously analyzed as part of a large international sequencing project.

In the original analysis, de novo mutational signature extraction had captured 288 distinct mutation categories. The resulting patterns had been decomposed into combinations of known reference signatures using an optimized method that excluded signature types considered biologically implausible for this tumor type. One extracted pattern, a single base substitution named SBS_D, achieved only a borderline fit during that reconstruction.

For the current study, the team reexamined SBS_D using several independent approaches. They compared its mutation pattern against a related signature extracted from the same cohort that closely matched a known reference signature associated with reactive oxygen species damage. Using copy number and variant-calling data, the researchers timed when the mutations attributed to each signature occurred during tumor evolution and classified them as early or late relative to chromosomal gains.

They also examined where the mutations occurred relative to nucleosomes, which are the protein structures around which DNA is wound. To confirm reproducibility, the researchers performed independent signature extraction on three separate CRC cohorts from the United States, Sweden, and the United Kingdom, using only treatment-naive, microsatellite-stable, repair-proficient samples.

They examined whether SBS_D activity was related to clinical features such as tumor stage and driver gene mutations using regression models, and screened for pathogenic mutations in DNA repair genes.

Furthermore, they conducted differential gene expression and pathway enrichment analyses on samples with available ribonucleic acid (RNA) sequencing data. Finally, the team applied an alternative decomposition approach permitting the full reference signature catalog, without exclusions, to reassess how SBS_D was linked to other known mutational processes.

Major Findings

The study found that SBS_D represents a mutational process distinct from the known signature it had previously been folded into, and that this pattern is reproducibly detected across multiple independent CRC cohorts.

Analysis of mutation timing showed that SBS_D mutations tended to accumulate late in tumor development, while the related known signature appeared consistently from early stages onward. The two patterns also diverged in their spatial relationship to nucleosomes, with SBS_D mutations being notably depleted in the DNA wrapped around these structures, while the known signature clustered near the nucleosomes. This distinction was considered unusual, since only a small number of previously studied mutation signatures have shown such depletion.

Moreover, when the researchers used an unrestricted decomposition approach, SBS_D appeared to resemble signatures typically linked to a faulty DNA polymerase enzyme and to defective mismatch repair, both of which are established drivers of hypermutation in tumors.

Interestingly, tumors carrying SBS_D lacked the elevated mutation burden and genomic instability patterns that are characteristic of repair-deficient cancers. This observation hinted against a straightforward repair-defect explanation. The signature was, however, associated with a modestly higher burden of small insertions and deletions, and its activity correlated with a specific indel pattern previously linked to replication slippage, hinting at a subtler connection to DNA repair fidelity rather than outright repair failure.

Because SBS_D contributes a meaningful share of mutations, roughly a sixth of all mutations in tumors where it is active, its consistent presence across cohorts from different countries suggests it reflects a genuine and previously overlooked process in colorectal tumor biology. However, the researchers noted that the gene expression differences associated with SBS_D were modest and should be interpreted cautiously, and that the precise molecular origin of the signature remains unresolved, warranting further experimental investigation.

Conclusions

This study identified and validated a previously overlooked mutational signature in CRC, now provisionally designated as SBS111 in the COSMIC database. By combining mutation timing, genomic positioning, and analysis across independent cohorts, the researchers demonstrated that this pattern reflects a distinct process rather than a mixture of known signatures.

Although its precise biological cause remains uncertain, the findings suggest a link to DNA repair infidelity and highlight how borderline signatures in cancer genomes can be missed.