When vets told a dog owner his pet had little time left, he turned to AI. Paul Conyngham was a data engineer with no background in biology. After surgery and chemotherapy failed, he used AI tools, DNA sequencing of the tumour and support from a university lab to help design a custom cancer vaccine. The treatment was not a cure. The University of New South Wales reported that the tumour shrank and the dog’s quality of life improved.
He started by using AI to work out what might be possible. Then he had the tumour’s DNA sequenced, which turned the cancer from tissue into data. Comparing the dog’s normal DNA with the tumour’s let him focus on the mutations most likely driving the disease. Instead of searching blindly, he could narrow in on a treatment path built around a single case.
AI sped up the discovery work that had to happen before the lab and its researchers could act. Part of that process used AlphaFold, an AI system from Google DeepMind that predicts the shapes of proteins. It modelled the mutated proteins and narrowed down the possible targets. Researchers then turned that analysis into a personalised mRNA vaccine, built from genetic instructions specific to the cancer.
The harder barrier was the system around the science. Conyngham spent hours each day for three months working through ethics approval before the vaccine could be given. Even with discovery moving fast, approval and coordination still set the pace.
Conyngham’s case has clear limits, since it involved one dog, one tumour profile and one response. It was not a formal trial, and it does not prove a general cure. The same core idea, though, is already being tested in people. In pancreatic cancer, vaccines tailored to each patient’s tumour stimulated immune activity in a small human study. In melanoma, a personalised therapy given with the immunotherapy drug pembrolizumab lengthened the time before the cancer returned, compared with pembrolizumab alone.
The step AI shortened here, from raw biological data to a treatment idea experts can test, is where a large share of early discovery cost sits. One motivated person with the right tools got far enough through it to bring university researchers and specialists into his dog’s treatment.
