Zhixiong Lin, head of AI Drug Discovery at XtalPi, presented on the acceleration of drug discovery using AI and next-generation automation. He began by introducing XtalPi, a science and research-driven company with four global sites and around 700 employees, 70% of whom were part of the science and tech team. The company aimed to develop a new paradigm for future drug discovery through various platforms, including small molecule drug discovery and solid-state platforms.
Lin highlighted XtalPi's collaboration with Pfizer in 2018 for solid-state prediction and their role in accelerating the process of Pfizer's COVID-19 drug, Paxlovid, in 2022. He also mentioned their partnership with Lilly in 2023, which received positive feedback. The core of Lin's presentation focused on XtalPi's AI drug discovery platform. He explained that their vision was to build a new paradigm for future drug discovery by designing smarter using AI algorithms and making faster using automation. The goal is to improve the success rate of the drug discovery process and bring effective medicine to patients more quickly.
Lin provided a case study to demonstrate the strength of their AI platform. The case involved a collaboration with a partner who had a flexible target that introduced specific challenges for computational modelling. XtalPi generated a million compounds using AI models, filtered them down to 100,000, and assessed 5,000 compounds in a single month using their XFEP platform. They selected 30 compounds for synthesis, and 20 of those compounds were made, with 40% showing high potency.
Lin stressed how combining AI with physics-based modelling can achieve an accurate and efficient exploration of chemical space. He noted that while AI models were powerful, they had limitations, and physics-based modelling could complement AI by providing more accurate binding affinity predictions.
In the final part of his presentation, Lin touched on XtalPi's automation efforts, which began in 2020 and had grown to include 200 workstations by 2023. The automation platform aimed to achieve rapid library synthesis with superior novelty and diversity, emphasising human-machine interactions and data protection. Lin concluded by highlighting the potential for synergy between XtalPi's chemistry expertise and partners strong in biology.
