Researchers at the University of Washington have developed a generative AI tool capable of designing functional cancer antibodies from scratch. This breakthrough could significantly accelerate the development of new antibody drugs and transform the $200 billion antibody drug industry.
The AI, trained on images of antibodies bound to their targets, can generate novel antibody designs expected to bind to proteins on various targets, including viruses, bacteria, and cancer cells. The team focused on engineering the six protein loops on the antibody's arms that bind to the target. Lab tests confirmed that the AI-designed antibodies bound to their intended targets as predicted.
This new approach offers a potentially faster and cheaper alternative to traditional antibody development methods, which involve immunising animals or screening large databases. The technology could lead to more effective and personalised treatments for cancer, autoimmune diseases, and viral infections.
