Artificial Intelligence and Gene Therapy to Accelerate Cures
By Dylan Morris
Today, we are proud to welcome Dyno Therapeutics out of stealth and into the CRV family. The company has announced not one, but two major strategic collaborations with leaders in the gene therapy space. We’re thrilled to support Dyno and their exciting work combining artificial intelligence and synthetic biology to provide gene therapies to patients who need them most.
Gene therapy aims to resolve underlying genetic issues that are the root cause of many diseases. More than 10,000 human diseases are causally mapped to specific variations in our genomes. There is now reason to believe that a majority of these diseases will be treatable, if not curable, with gene therapy. The first gene therapy was approved only two and a half years ago; the second was approved last year. Both were tied to multibillion-dollar acquisitions. Momentum in the field is extraordinary and should continue apace, as industry players compete for access to this game-changing new technology.
The first approvals were for rare diseases. Treatments for more common diseases, such as heart failure and sickle cell anemia, are in active clinical development. Three gene therapies for hemophilia are currently in Phase 3 trials. A first approval could come within the year. While the economic value of curing these and other diseases is utterly enormous, the human value is impossible to state. Once we know that a cure is possible, we have a moral obligation to patients and their loved ones to do everything we can to accelerate its development. For some of these patients, every day delayed is another day of suffering.
There are several areas where further innovation will move the field forward faster. Key among these is delivering gene therapies to the right location in the body. This is a decidedly difficult problem and we currently borrow from nature to solve it. Viruses have evolved precisely to this purpose — to deliver genetic material into cells. Their replication instructions, in the form of DNA or RNA, are packaged within self-assembling protein shells called capsids. The physical and chemical properties of these capsids determine the types of cells they infect.
Most gene therapies repurpose the innocuous adeno-associated virus (AAV) capsid to deliver corrective genes to cells. The field has been limited to naturally occurring variants and the few cell types they infect. To improve delivery to other cell types, one could introduce random changes to the capsid protein and test for better variants. However, the space of possible AAV capsids is unreasonably large (20⁷³⁶). Random sampling is a woefully inadequate search strategy for a space this size. Vast regions of functional design space will never be explored.
Dyno overcomes this issue brilliantly — integrating artificial intelligence with synthetic biology to engineer improved AAV capsids. Some of the breakthroughs that enable Dyno’s approach were recently described in a research report in the journal Science. These tools enable massively parallel in vivo experimentation. Sophisticated machine learning algorithms inform the search through capsid design space. Machine-guided design iteratively drives systematic mutagenesis towards desired phenotypes, such as delivery to specific tissues. Dyno’s approach is an exquisite demonstration of how the best engineers think about and solve biological problems.
Dyno’s CEO and co-founder, Eric Kelsic, is an immensely talented entrepreneur who has been thinking about biology for over 15 years. Eric received his BS in physics from Caltech, where he got his intro to the wet lab with synthetic biology. From there, Eric went to Harvard where he completed his PhD in systems biology before doing a postdoc with George Church, master of multiplexing, whose many contributions to the development of DNA sequencing and synthesis technology cannot be underrated. George and his lab show us the edge of the possible.
Dyno’s team blends expertise across gene therapy and artificial intelligence. The company has hired some of the world’s best machine learning scientists and AAV engineers. Together, we are building a foundational company to provide improved delivery vectors for the entire gene therapy space. We are inspired in our quest to bring more cures to more people, faster. If this inspires you, too, get in touch. Dyno is hiring for key roles in software, biology, and business.