104437Computational Antibody Design with Machine LearningClosedFeasibility StudiesInnovate UK"Due to their high specificity and low toxicity, antibodies are attractive candidates for improved, more efficient, easier to deliver, and safer therapeutics. In 2016, 5/10 best-selling drugs worldwide were antibodies, including the record-breaking Humira which took £12bn in sales, the highest ever for any pharmaceutical. The global antibody therapeutics market is expected to continue to grow at a remarkable pace, reaching £88bn by 2020, driven by the high incidence of chronic diseases which urgently require more effective treatments. High R&D costs and strict regulatory measures may however remain the longstanding challenges for pharmaceutical companies, hampering market growth. Antibody discovery is the process of discovering an antibody to bind to a particular antigen. Existing methods are through immunisation of animals (rabbits, mice), or by screening a large antibody library. While these physical platforms are usually able to discover binding antibodies, they require targets that are well-formed and available purified in sufficient quantities. Even then, the process is time consuming and expensive, and existing techniques suffer from throughput, scalability, repeatability and quality issues. Antiverse is building a world-first computational antibody discovery platform combining state-of-the-art machine learning techniques to predict antibodies that bind to a given antigen target with high affinity. The company has a vision to overturn the £3.5bn antibody drug discovery market, replacing existing antibody discovery techniques with a novel solution to design antibody drugs _in silico_, whilst massively reducing the cost (up to £300K/candidate) and time to discover new candidates, from 3 to 18 months down to just one day. Antiverse is well placed to exploit this opportunity, having already established a basic model with an initial dataset containing public data to prove technical feasibility. The company has secured laboratory space for generating a proprietary dataset at scale to develop the model and has good industry links and connections with major biopharmaceutical companies and contract research organisations (CROs) that have already expressed a keen interest in trialling the solution once developed. This project will enable Antiverse to develop the machine learning algorithm and a large proprietary dataset with at least 10,000 data points required to prove the generative model, facilitating the subsequent development of a platform that can be trialled with industry partners to accelerate the service offering towards commercialisation by mid-2020, helping to establish this UK SME at the forefront of the antibody drug discovery market poised for significant growth."