Key Takeaways
- A surprise discovery at soybean-genetics firm Confluence Genetics produced Cas-CLEAR, a CRISPR platform now being explored as a potential CRISPR cancer treatment that destroys a tumor cell’s DNA and RNA while leaving healthy cells intact.
- Researchers at the University of Utah and Jennifer Doudna’s lab at UC Berkeley have shown Cas-CLEAR working across several cancer types and, most recently, in a mouse model — though Begemann cautions it has not yet cured cancer in a mouse.
- Confluence intends to license the technology rather than build drugs itself — an “AWS of CRISPR” — partnering with oncology and delivery specialists to turn the platform into an actual treatment.
- The company’s core business is premium soybean genetics: varieties that deliver new value toaquaculture, human food, dairy, and poultry markets and earning farmers premiums above commodity soybeans..
- Its CropOS AI platform pairs sequence and phenotype data with proprietary algorithms to predict protein, oil, and carbohydrate profiles, cutting development of new soybean varieties from eight-to-ten years to roughly six.
Listen or Watch The Entire Episode Below
The most surprising thing Confluence Genetics is working on has almost nothing to do with soybeans. Inside a company built to breed a better bean, a surprise discovery has produced what could one day become a CRISPR cancer treatment — a way to destroy a tumor’s cells while leaving the healthy ones around them untouched. It is an unlikely place for an oncology idea to take root, and the scientist leading the work is the first to say it is still early.
Matt Begemann grew up on the receiving end of the product his company improves. His family farms soybeans in Missouri, and after a biochemistry degree at the University of Missouri and a PhD in microbiology from the University of Wisconsin, he spent about a decade at Benson Hill working on trait discovery, gene-editing tools, and AI for plant breeding. When Confluence Genetics acquired Benson Hill about a year ago, that work — and the surprise discovery that would point toward a cancer application — came with it.
A Surprise Discovery That Pointed to a CRISPR Cancer Treatment
The platform is called Cas-CLEAR, and Begemann is careful to draw a line others might blur: it is a CRISPR platform, not a gene-editing tool. Conventional CRISPR uses a guide RNA to steer a nuclease to a precise spot in the genome and make a single cut. Cas-CLEAR does something else. Similar to other CRISPR systems, it uses a guide RNA to find the target – in this case it is an RNA target. After it precisely finds its initial target, the nuclease changes shape and then “indiscriminately chews up the DNA and the RNA of that cell,” leading to targeted cell death.
The team first noticed the effect when the bacteria they were testing these new CRISPR nucleases in kept dying once the nuclease was turned on. Rather than treat that as a failure, they pivoted: if it kills a cell that precisely, it could work as an antimicrobial. Cas-CLEAR keeps the single-base-pair targeting that makes CRISPR valuable, but turns it toward targeted cell elimination rather than editing. It can kill bacteria; it can kill a higher organism like yeast. And that raised a much larger question — one that leads straight to oncology.
Testing the CRISPR Cancer Treatment on Cancer Cells
Cancer, Begemann notes, is fundamentally a set of oncogenic mutations, and modern precision medicine is increasingly good at sequencing a patient’s tumor to identify exactly which ones. The concept behind a Cas-CLEAR CRISPR cancer treatment is to design the platform for that specific tumor and deliver what he describes as “cell-specific chemotherapy” — destroying the cancer cells while leaving healthy cells intact. The company has seen it work in its own lab, and Begemann says outside groups, including researchers in the Liu lab at the University of Utah and Jennifer Doudna’s lab at UC Berkeley, have shown it working across several cancer types, including in a mouse model.
“Not everybody knows a farmer. Everybody eats, but not everybody knows a farmer. And everybody’s been impacted by cancer.”
He is quick to add the caveat that matters: the platform works in cell cultures and functions in a mouse model for lung and liver cancer, but “we haven’t fully cured cancer in a mouse yet.” Turning it into a real CRISPR cancer treatment will take more work on the underlying chemistry and on the delivery vehicle needed to carry it safely into the body.
Why the CRISPR Cancer Treatment Stays a Licensing Play
Confluence has no intention of becoming a drug company. Begemann and most of the staff remain focused on soybean genetics; only a small team works on the CRISPR chemistry underneath Cas-CLEAR. The plan is to be a technology provider — licensing the platform and striking co-development deals that pair Confluence’s expertise in CRISPR chemistry with partners who bring the delivery mechanisms and disease-specific knowledge. It is an “AWS of CRISPR” model, an analogy raised in the interview that Begemann was happy to accept: make the core technology excellent, then let others build the products on top of it. For a cancer treatment, that means handing the asset to an oncology partner once the platform is ready — and collecting a royalty on the back end.
It Started With Soybeans
For all the attention the cancer work draws, it grew out of a business that is squarely agricultural. “Soybeans, at the end of the day, are the most efficient source of protein in the world,” Begemann says, and the world’s shortfall is not where people often assume: “The world doesn’t seem to be short calories, but we are short protein.” Confluence breeds premium-quality soybean varieties that let growers earn more — beans bred for aquaculture, human food, dairy, and poultry markets that pay above commodity prices. That matters to farmers under constant pressure. “As a soybean grower, you’re gambling with your net worth every single year,” he says, and the company’s pitch is a seed that earns a premium without asking farmers to change how they operate.
Data as the Real Moat
The durable advantage, Begemann argues, is data. Confluence’s current soybean portfolio is not gene-edited at all; its quality traits come from natural genetic variation gathered worldwide and combined through more than a decade of breeding. That history feeds an AI platform called CropOS, which integrates sequence and phenotype data to predict a variety’s yield and, especially, its protein, oil, and carbohydrate profiles. A breeder specifies what they want — a maturity group for a given region, high protein, low anti-nutrients for poultry — and CropOS runs millions of simulations to recommend which crosses to make and how to select the offspring, cutting development of a new variety from eight-to-ten years down to roughly six. The edge, he argues, is not the model. “Everybody’s got great models,” he says; the real differentiator is the strategic data behind them. “You’ve got to have data, and it’s got to be high-quality data, and it’s got to be connected.”
The Next Five Years
Asked where Confluence will be in five years, Begemann does not reach for the moonshot. The near-term work is squarely in soybeans: capturing more of the roughly 35 million acres of U.S. production that flow into domestic poultry & swine feed, growing the high-oleic oil business across human food and dairy, and launching herbicide-tolerant varieties that make adoption easier for farmers who would rather not overhaul their operations. The CRISPR work continues in parallel, with licensing as the goal — and if the platform’s plant-defense applications pan out, they feed back into the core soybean business and potentially into crops such as corn, potatoes, and cotton. Food and human health, in his telling, are linked, and a technology that can improve both is worth building carefully. Those interested in the agricultural side can start at Confluence.ag; inquiries about the platform go to cas-clear.com.
