In early 2024, Sydney tech entrepreneur Paul Conyngham faced devastating news: his beloved rescue dog Rosie, an eight-year-old Staffy-Shar Pei mix adopted from bushland in 2019, had been diagnosed with aggressive mast cell cancer. Veterinarians warned that the tumors—some as large as a tennis ball—were incurable with standard treatments, giving her only months to live. Traditional chemotherapy and surgery slowed the disease but failed to shrink the masses, leaving Conyngham, who had no formal background in biology or medicine, desperate to find another path for his "best mate" who had stood by him through tough times.
Refusing to accept the prognosis, Conyngham leveraged his expertise as a data scientist and machine learning engineer. He paid approximately $3,000 to have Rosie's healthy DNA and tumor tissue sequenced at the University of New South Wales (UNSW) Ramaciotti Centre for Genomics. Turning to accessible AI tools, he used ChatGPT extensively to brainstorm treatment strategies, identify relevant research on immunotherapy, and guide the analysis of the massive genomic dataset. The AI suggested focusing on neoantigens—unique mutations in the cancer cells that could train the immune system to attack them—while Google's AlphaFold helped model the structures of mutated proteins like those involving c-KIT, a common driver in mast cell tumors.
Conyngham's process involved sifting through hundreds of gigabytes of genetic data to pinpoint promising targets for a personalized mRNA vaccine, similar in concept to those developed for COVID-19 but tailored specifically to Rosie's tumor mutations. Reports indicate he collaborated with experts, including nanomedicine pioneer Prof. Pall Thordarson at UNSW's RNA Institute, who manufactured the custom mRNA lipid nanoparticle vaccine in the lab based on the AI-assisted design. Some accounts note that X's Grok AI contributed to refining the final vaccine sequence construct. The experimental treatment, a one-off product, received ethics approval and was administered at a veterinary research center, with booster doses following the initial injection around late 2025.
The results have been remarkable though not miraculous. Within weeks, the large tumor on Rosie's leg shrank dramatically—reports cite reductions of about 50-75 percent—while other tumors also decreased in size. Rosie regained her energy, jumping fences to chase rabbits and enjoying a significantly improved quality of life far beyond the initial grim outlook. Scientists at UNSW described Conyngham's genomic analysis as impressive for a non-specialist, highlighting how AI tools democratized complex data interpretation. This appears to mark the first known personalized mRNA cancer vaccine designed for a dog.
Experts emphasize important caveats: this remains a single anecdotal case, not a proven cure or broadly validated therapy. Oncologists have urged caution, noting that long-term safety, efficacy, and potential risks of experimental mRNA treatments in veterinary medicine require further study through controlled trials. Conyngham combined AI with rigorous human oversight, including ethics paperwork and collaboration with university labs—elements essential to the outcome. Viral headlines sometimes overstate the role of AI or misattribute tools (such as claiming primary use of Claude), but the core story underscores AI's supportive power in accelerating personalized approaches.
The case of Rosie and Paul Conyngham points toward a future where AI lowers barriers in medicine, enabling motivated individuals to explore innovative solutions alongside expert teams. Conyngham has shared details of the protocol openly, aiming to help others facing similar challenges in canine cancer. While far from replacing traditional veterinary care or human oncology research, it illustrates the potential of mRNA technology and AI-assisted design in advancing individualized treatments for both animals and, eventually, people. Rosie's ongoing battle continues, but her story offers hope and a glimpse into emerging capabilities in the fight against cancer.
Refusing to accept the prognosis, Conyngham leveraged his expertise as a data scientist and machine learning engineer. He paid approximately $3,000 to have Rosie's healthy DNA and tumor tissue sequenced at the University of New South Wales (UNSW) Ramaciotti Centre for Genomics. Turning to accessible AI tools, he used ChatGPT extensively to brainstorm treatment strategies, identify relevant research on immunotherapy, and guide the analysis of the massive genomic dataset. The AI suggested focusing on neoantigens—unique mutations in the cancer cells that could train the immune system to attack them—while Google's AlphaFold helped model the structures of mutated proteins like those involving c-KIT, a common driver in mast cell tumors.
Conyngham's process involved sifting through hundreds of gigabytes of genetic data to pinpoint promising targets for a personalized mRNA vaccine, similar in concept to those developed for COVID-19 but tailored specifically to Rosie's tumor mutations. Reports indicate he collaborated with experts, including nanomedicine pioneer Prof. Pall Thordarson at UNSW's RNA Institute, who manufactured the custom mRNA lipid nanoparticle vaccine in the lab based on the AI-assisted design. Some accounts note that X's Grok AI contributed to refining the final vaccine sequence construct. The experimental treatment, a one-off product, received ethics approval and was administered at a veterinary research center, with booster doses following the initial injection around late 2025.
The results have been remarkable though not miraculous. Within weeks, the large tumor on Rosie's leg shrank dramatically—reports cite reductions of about 50-75 percent—while other tumors also decreased in size. Rosie regained her energy, jumping fences to chase rabbits and enjoying a significantly improved quality of life far beyond the initial grim outlook. Scientists at UNSW described Conyngham's genomic analysis as impressive for a non-specialist, highlighting how AI tools democratized complex data interpretation. This appears to mark the first known personalized mRNA cancer vaccine designed for a dog.
Experts emphasize important caveats: this remains a single anecdotal case, not a proven cure or broadly validated therapy. Oncologists have urged caution, noting that long-term safety, efficacy, and potential risks of experimental mRNA treatments in veterinary medicine require further study through controlled trials. Conyngham combined AI with rigorous human oversight, including ethics paperwork and collaboration with university labs—elements essential to the outcome. Viral headlines sometimes overstate the role of AI or misattribute tools (such as claiming primary use of Claude), but the core story underscores AI's supportive power in accelerating personalized approaches.
The case of Rosie and Paul Conyngham points toward a future where AI lowers barriers in medicine, enabling motivated individuals to explore innovative solutions alongside expert teams. Conyngham has shared details of the protocol openly, aiming to help others facing similar challenges in canine cancer. While far from replacing traditional veterinary care or human oncology research, it illustrates the potential of mRNA technology and AI-assisted design in advancing individualized treatments for both animals and, eventually, people. Rosie's ongoing battle continues, but her story offers hope and a glimpse into emerging capabilities in the fight against cancer.
