- Pfizer and CeMM collaborated for three years on an AI-driven drug discovery method.
- The method identifies small molecules binding to human proteins, expediting drug development.
- CeMM researchers created a catalog of interactions that are freely available via a web app.
- Over 90% of marketed drugs are small molecules; only 20% of human proteins have known ligands.
- The chemical proteomics approach tested 407 small molecule fragments against human proteins, revealing nearly 47,700 interactions.
- Machine learning framework predicts small molecule-protein binding and accessibility.
- Pfizer, an early AI adopter, funded the research and has previously used AI in vaccine and drug safety monitoring and in developing Paxlovid for COVID-19.
Main AI News:
In a landmark partnership spanning three years, Pfizer and the Research Center for Molecular Medicine of the Austrian Academy of Sciences (CeMM) have forged a pioneering path in drug discovery. Their collaboration has birthed a cutting-edge AI-driven methodology poised to revolutionize the identification of small molecules with therapeutic potential, as unveiled in an article published in Science on April 25.
CeMM researchers, led by Dr. Georg Winter, have masterfully crafted and scaled an AI and machine learning platform. This innovative system meticulously evaluates the interactions between hundreds of small molecules and thousands of human proteins. The result? A comprehensive catalog that serves as a springboard for novel drug development endeavors. Notably, all models and data are generously shared with the global research community through a user-friendly web application, as highlighted in a press release from CeMM.
Dr. Winter expressed his astonishment at the transformative power of artificial intelligence and machine learning in unraveling the intricacies of small-molecule behavior within human cells. He emphasized the potential of their catalog and AI models to streamline drug discovery processes significantly.
The significance of this breakthrough cannot be overstated, given that over 90% of all marketed drugs are small molecules that exert their therapeutic effects by binding to proteins. However, the identification of ligands for merely 20% of human proteins has long been a bottleneck in drug development and medical research at large. Efforts to address this gap, such as the recent dataset published by scientists at Paris’ Institut Pasteur and the creation of tools like Pharos by the Illuminating the Druggable Genome consortium, underscore the urgency and importance of innovative approaches.
CeMM and Pfizer’s methodology revolves around “chemical proteomics,” employing chemical probes to investigate the binding interactions of small molecule compounds with proteins of interest. Their exhaustive experimentation involving a library of 407 small molecule fragments yielded a treasure trove of nearly 47,700 distinct protein-ligand interactions across over 2,600 unique proteins. Notably, close to 90% of these proteins lacked previously known ligands, marking a significant advancement in our understanding.
The translational potential of this approach is exemplified by its application in developing synthetic ligands targeting various proteins. Furthermore, the data gathered served as the foundation for a sophisticated machine learning framework. This framework can predict the binding affinity of small molecules to diverse proteins and assess the accessibility of these proteins. Additionally, it can anticipate interactions with specific subsets of proteins, such as RNA-binding or transporter proteins, as well as proteins localized within specific cellular compartments.
Pfizer’s early involvement and financial support underscore its commitment to leveraging AI in drug development. Having embraced this technology for monitoring vaccine and medicine safety since 2014, Pfizer’s pioneering spirit culminated in the development of Paxlovid, a breakthrough antiviral for COVID-19. This collaborative endeavor with CeMM represents yet another milestone in Pfizer’s journey towards innovation and scientific excellence.
Conclusion:
Pfizer’s collaboration with CeMM marks a significant advancement in drug discovery, offering a transformative approach that addresses longstanding challenges in identifying therapeutic targets. This breakthrough not only accelerates the development of novel treatments but also underscores the pivotal role of AI in shaping the future of pharmaceutical innovation. Pharmaceutical companies must embrace such collaborative efforts and leverage AI technologies to stay competitive and drive meaningful advancements in healthcare.
