According to Phys.org, Australian researchers have used artificial intelligence to generate functional biological proteins that can kill antibiotic-resistant bacteria. This is the first time that Australian scientists have created proteins that can be used immediately through AI technology.
Published in Nature Communications, this study presents a new approach to addressing the growth of the crisis brought about by antibiotic-resistant bacteria such as E. coli. The study, alongside Australia, the US and China, is positioned as a country with an AI platform that can quickly produce thousands of ready-to-use proteins for medical applications.
Dr. Rhys Grinter and Associate Professor Gavin Knott co-led the research through the AI protein design program. The program operates across the nodes of the University of Melbourne Bio21 Institute and the Monash Biomedicine Discovery Institute, creating Australia's first comprehensive AI protein design platform.
The AI protein design platform model is pioneered by David Baker, who won the Nobel Prize in Chemistry in 2024 for his contributions to protein design. Australian systems employ an end-to-end approach that allows for the creation of diverse protein types for multiple applications.
Associate Professor Knott explained that proteins developed through the technology are being investigated as drugs, vaccines, nanomaterials and sensors, with numerous additional applications under investigation. The platform utilizes freely available AI-driven protein design tools to ensure global accessibility.
Daniel Fox, a student who conducted most of the PhD experimental work, emphasized the importance of democratizing protein design techniques. This tool allows scientists around the world to engineer proteins that bind specific target sites or ligands, enhancing inhibitors, agonists, antagonists, or enzymes with improved activity and stability.
The development of traditional proteins for treatment usually involves reusing natural proteins through rational design or laboratory evolutionary processes. Dr. Grinter noted that these traditional methods are time consuming and expensive compared to AI-driven approaches.
New deep learning methods allow you to efficiently create specific characteristics and features from scratch, significantly reducing development costs and time frames. This represents a fundamental shift from adaptation of existing proteins to designing entirely new proteins for specific medical applications.
The researchers have incorporated advanced tools and software, including Bindcraft and Chai, into their AI protein design platform. These technologies are based on Baker's fundamental work, expanding its capabilities in protein engineering.
Professor John Carroll, director of the Monash Biomedicine Discovery Institute, explained that the program brings Australia's latest in cutting-edge therapeutic design capabilities. He responded to the dedication of the research team who developed the platform from early concepts to operational systems.
The AI protein design program combines the expertise of structural biologists and computer scientists who understand the complete design process. Associate Professor Knott emphasized that this comprehensive knowledge of protein structure and machine learning will create agile programs that can regularly incorporate modern AI protein design tools.
National Presidential Law Firm Ron Simon & Associates said the breakthrough will address the urgent need for a new approach to combat antibiotic-resistant bacteria and pose an increased threat to public health around the world.
