An Australian ML researcher, used ChatGPT+AlphaFold to shrink 75% of his life-threatened dog’s MCT cancerous tumor, developing a personalized mRNA vaccine in just two months - after sequencing his dog’s DNA for $2,000

Paul Conyngham, an Australian data analyst with machine learning experience but no formal biology background, successfully developed a personalized mRNA vaccine for his dog Rosie's aggressive mast cell tumor using publicly available AI tools. The process began with Conyngham paying approximately $2,000 to sequence the tumor's DNA through the University of New South Wales' Ramaciotti Centre for Genomics, where computational biologist Martin Smith initially expressed skepticism about the computational burden. Conyngham assured Smith he could handle the analysis, then used ChatGPT to identify tumor-specific neoantigens—mutated proteins that could trigger an immune response—and employed Google DeepMind's AlphaFold to predict the three-dimensional structures of these proteins.

After identifying promising vaccine targets through this AI-driven analysis, Conyngham collaborated with UNSW RNA Institute expert Pall Thordarson to synthesize the mRNA vaccine from a DNA template. The resulting personalized treatment, developed in just two months, achieved remarkable results: Rosie's tumor shrank by approximately 75%, transforming a life-threatening cancer into a manageable condition. Thordarson emphasized the breakthrough nature of this achievement, noting that Conyngham accomplished what traditionally requires years of specialized training by leveraging AI tools to bridge knowledge gaps in genomics, immunology, and structural biology.

This case demonstrates how accessible AI technologies are lowering barriers to complex biomedical research, enabling non-specialists to contribute to therapeutic development. While still requiring expert collaboration for synthesis and validation, the combination of large language models like ChatGPT and protein prediction tools like AlphaFold created a workflow that dramatically compressed the vaccine design timeline. The success suggests similar approaches could accelerate development of personalized treatments for human cancers, potentially reducing costs and development cycles for targeted therapies that address individual genetic profiles.