60% of therapeutics are antibody-based. The discovery process starts with the identification of a target or antigen, which companies must find a matching antibody for through trial and error: the testing of “libraries” of millions of antibodies against the target to see what “binds”. Because no library contains all possible antibodies, this process is not guaranteed to work, and firms must create larger and larger libraries, which take longer and cost more. The average process is longer than 18 months and costs hundreds of thousands of pounds. Being able to explore the entire possibility space and conduct tighter iteration cycle from target to drug would allow scientists to quickly validate target hypotheses.
Antiverse is the first company in the world capable of developing antibody drug candidates in one day for any given antigen. It was previously thought that this was impossible, owing to the physical size and complexity of antibodies. The team bring to bear an unlikely combination of robotics, state of the art machine learning and cell-free protein synthesis. Their product can take an antigen target sequence provided by the customer and conduct high-throughput screening of all possible antibody sequences to detect the sequences which produce a high-affinity antibody for the target: in a single day, thus reducing the time by up to 18 months. Within their first 6 months of operation they have achieved a commercial partnership and developed a beta model which operates at 99.5% accuracy in predicting binding affinity between a given antibody and a target antigen.
“This will put us out of business.” – Antibody-focused CRO