The process of creating new medications has long been a complex and costly endeavor, often fraught with obstacles that can lead to failure. The journey from concept to patient is lengthy and expensive, with many potential drug candidates falling by the wayside. This challenge is even more pronounced in the realm of biologic medicines, which are derived from engineered proteins and are crucial for treating various acute and chronic diseases. Scientists must sift through vast libraries of molecules to identify the few that can effectively bind to the appropriate targets, maintain stability in the human body, and be manufactured at scale.
In recent years, artificial intelligence (AI) has emerged as a transformative tool in pharmaceutical research and development (R&D), significantly accelerating the drug discovery process. Companies like AstraZeneca are leveraging AI to enhance their biologic drug development efforts, assembling specialized engineering teams to harness this technology. According to Puja Sapra, Senior Vice President and Head of R&D for Biologics Engineering and Oncology Targeted Discovery at AstraZeneca, every aspect of the drug development cycle—designing, testing, and analyzing—is now enhanced by computational capabilities. This shift results in shorter cycle times and increased productivity, enabling scientists to focus on the most promising candidates through a streamlined build-measure-learn approach.
Beyond improving efficiency, AI is paving the way for the discovery of innovative classes of medicines. Traditional biologics typically target single disease pathways, but the next wave of drugs may effectively engage multiple targets or deliver therapeutic agents to specific cells. This necessitates sophisticated optimization across a multitude of variables. AI-driven models can assist in identifying which targets to prioritize and optimizing parameters such as potency, stability, manufacturability, and safety. As Sapra notes, the concept of “drugging the undruggable” is becoming a reality, with technologies that could enable the development of treatments for previously insurmountable targets.
To maximize the potential of AI in drug design, access to high-quality biological data is crucial. AstraZeneca is actively generating diverse datasets through extensive experimental trials, which inform and refine their AI models. The company is also constructing a state-of-the-art ‘lab of the future’ in Cambridge, Massachusetts, where AI and robotic automation will work hand-in-hand to create a continuous discovery loop. This innovative facility will allow for rapid experimentation and data generation, significantly outpacing traditional workflows.
Looking ahead, the ultimate goal is achieving de novo design, where AI autonomously generates entirely new protein sequences tailored for specific therapeutic properties. While substantial progress has been made, achieving this vision will require standardized training data, robust evaluation benchmarks, and interdisciplinary teams skilled in the intersection of machine learning and biology. As AstraZeneca tackles the complexities of safety prediction through virtual clinical trials and advanced modeling techniques, the future of medicine development appears to be on the brink of a revolutionary transformation, promising remarkable benefits for patients worldwide.
Source: How AI helps scientists design the next generation of medicines via MIT Technology Review