Sydney Tech Entrepreneur Uses AI to Develop Custom mRNA Cancer Vaccine for Dying Dog
Key Takeaways
- ▸AI tools like ChatGPT and AlphaFold enabled a non-medical professional to collaborate with researchers on developing a personalized cancer treatment
- ▸The custom mRNA vaccine represents the first personalized cancer vaccine designed for a dog and demonstrates the potential of precision medicine approaches
- ▸While not a cure, the treatment has produced significant clinical improvement and extended both quantity and quality of life for the patient
Summary
Sydney tech entrepreneur Paul Conyngham leveraged artificial intelligence tools to develop a personalized mRNA cancer vaccine for his dog Rosie after traditional treatments failed. After receiving a cancer diagnosis, Conyngham used OpenAI's ChatGPT to explore immunotherapy options and was directed to the University of New South Wales Ramaciotti Center for Genomics. He then employed Google DeepMind's AlphaFold to identify mutated proteins as treatment targets, and collaborated with UNSW nanomedicine pioneer Pall Thordarson to design a bespoke mRNA vaccine in under two months.
Rosie received her first vaccine injection in December 2024 and a booster in February 2025, with dramatic results: most of her tumors have shrunk significantly and her overall health and energy have improved markedly. Thordarson noted this marks the first personalized cancer vaccine designed for a dog and demonstrates that mRNA technology can democratize the process of designing cancer vaccines. While cautioning that some tumors remain and a full cure is not guaranteed, Conyngham reports the treatment has bought Rosie considerably more time and quality of life, enabling her to resume normal activities like chasing rabbits at the dog park.
- The case illustrates how AI technology could democratize and accelerate the development of personalized medical treatments for both animals and humans
Editorial Opinion
Rosie's story represents a compelling proof-of-concept for AI-assisted personalized medicine, demonstrating how large language models and protein-folding AI can democratize complex scientific research and compress timelines for breakthrough treatments. However, it's important to note this remains an experimental case with mixed results—while tumors shrank significantly, some remained unresponsive. The real significance lies not in overstating Rosie's recovery as a cure, but in validating the approach's potential to make precision oncology more accessible and rapid, setting the stage for human applications in cancer immunotherapy.
