"CRISPR-Cas13b Gene Fusion Silencing: A Programmable Way to Switch Off 1 in 5 Cancers"

Cancer has a habit of hiding its instructions in plain sight. A new study from Peter Mac researchers, published in Science Advances on October 19, 2026, takes aim at one of its most common tricks: gene fusions, the accidents that occur when DNA breaks and is repaired incorrectly, stitching parts of two different genes together. The fused gene can then drive a cell toward cancer, and the researchers have shown, in laboratory cell models, that a reprogrammed CRISPR tool can recognise the exact join and silence it.

The numbers explain why this matters beyond a single tumour type. Gene fusions are present in about 20 percent of all cancers, and the paper describes them as bona fide oncogenic drivers in 16.5 percent. Two examples carry the point. BCR-ABL1, the fusion of the BCR and ABL1 genes, is present in nearly all cases of chronic myeloid leukaemia. EWSR1-FLI1 is found in more than 85 percent of cases of Ewing's sarcoma, a rare cancer of bone and surrounding soft tissue. These are very different diseases, and that is precisely the attraction of the new approach: instead of building a separate drug for each fusion, the team reprogrammed one system to read different breakpoints.

The tool is CRISPR-Cas13b, and the distinction from the better-known Cas9 matters. Rather than cutting DNA, Cas13b is directed at the RNA transcript produced by the fused gene, recognising the breakpoint sequence where the two genes meet. That junction is a molecular fingerprint found only in cancer cells, in the researchers' framing, because healthy cells do not carry that fused transcript. In the proof-of-concept work, led by Dr. Wenxin Hu with Dr. Mohamad Fareh and colleagues, nine different gene fusions were successfully targeted and silenced in cell models, switching off their cancer-driving downstream effects.

What I find most interesting here is the shift in what "personalised" means. Personalised cancer medicine usually means matching a patient to an existing drug, or at best to an existing trial. This proposal inverts the order: sequence the patient's tumour, read its particular fusion breakpoint, then program the CRISPR system to match that tumour's fingerprint. Fareh puts it as a long-term vision of programmable cancer medicines tailored to the genetic abnormality driving an individual tumour. If that vision holds, the bottleneck moves from drug discovery to delivery and manufacturing — from inventing a new molecule to programming and shipping a new guide.

That "if" deserves to be stated plainly, because the study is candid about where it stops. The method has so far been demonstrated only in laboratory cell models. It has not been tested in animals or in people, and the next step is to determine how effectively the CRISPR machinery can be delivered to tumours inside the body, and whether it can selectively target cancer cells in more complex models. Delivery has humbled many elegant laboratory ideas before: a system that works in a dish still has to survive the bloodstream, reach the right tissue, enter the right cells, and do so without setting off the immune system or silencing the wrong transcripts. None of that is solved by a breakpoint sequence alone.

Still, there is a reason to take this seriously rather than file it as another promising cell result. The approach attacks a defined, shared mechanism rather than a single mutation, and it does so with a platform logic: change the program, not the machine. The same logic that lets one CRISPR system be pointed at nine fusions in the lab is what would, in principle, let it be pointed at a tenth, or at a fusion found in only a handful of patients, where a conventional drug would never be commercially viable. Rare fusions are exactly where platform approaches earn their keep, if the delivery problem yields.

So the honest summary is two-sided. On one side, a precise, programmable way to silence a class of cancer drivers that touches one in five cancers, demonstrated across nine fusions in cells. On the other, an untested delivery journey between a dish and a patient, with selectivity in a living body still to be proven. The study's contribution is not a treatment. It is a credible route to one, and the next result worth watching is not another cell line silenced, but the first animal in which the system reaches a tumour and works.