Despite the simplicity of its name, drug discovery broadly encompasses the rigorous process of identifying chemical compounds, also known as leads, which have the potential to be developed into a medical treatment. However, this process is a time-consuming, expensive investment that cannot guarantee success in later drug development phases. The reality of traditional lead discovery research is that for every novel wonder drug that enters the market, there are tens of thousands of candidate compounds which failed screening and were discarded. This high attrition rate of drug leads is a significant pain point for sponsors and contract research organizations (CROs), but could a solution be in sight with the introduction of AlphaFold? Read on to discover what this innovative technology means for the future of drug development!
The Dilemma of Protein Folding
Proteins are integral to supporting all functions of life at the cellular level. Although its specific sequence of amino acids is important for dictating how a protein will behave, the exact 3D structure formed plays an equally large role in protein function. The dilemma this has presented in biology research for decades is that scientists could not easily predict which shapes a protein’s primary structure might fold into. To date, methods such as nuclear magnetic resonance, X-ray crystallography, and cryo-electron microscopy have been tested with trial and error to determine protein structures. However, these are highly limited by inefficient manual work and the need for expensive specialized equipment. Using these traditional methods, the cost of determining crystal structures can quickly climb to tens or even hundreds of thousands of dollars.
What is AlphaFold?
Developed by DeepMind, AlphaFold is an open-access protein structure database built with artificial intelligence (AI) to predict a protein’s 3D structure. Its approach has helped inspire a computational race in the field of protein prediction, which now includes similarly effective AI tools such as ESMFold, OpenFold, RoseTTAFold, and more. To date, only a small fraction of the 200 million proteins we know of have had their functional shapes fully elucidated by scientists. AlphaFold and its competitors are helping close this knowledge gap by offering the ability to predict protein structures quickly and with scaled atomic accuracy. On July 28, 2022, DeepMind expanded the AlphaFold database from its initial catalog of 1 million proteins to include nearly all 200 million structures currently known to science.
One Step for AI, One Potential Leap for Drug Discovery?
AlphaFold addresses the challenge of predicting protein folding, but has this AI system changed the game for discovering new therapies? Although it offers a fairly reliable representation of what a protein may look like, pharmaceutical companies would not consider it a potential drug from its predicted structure alone. Based on this, AlphaFold’s introduction has not yet made a substantial impact on the drug discovery process, but it is still a worthwhile visualization tool. Revealing the physical properties of a target protein provides essential information about its role in disease pathophysiology. The more a sponsor’s or CRO’s R&D team understands the 3D structures of key proteins related to their indication of interest, the easier it becomes to design a potent ligand tailored for its target. For example, the μ-opioid receptor is critical for understanding the body’s response to opium derivatives, but this target has a particularly challenging crystal structure. AI-powered protein prediction technology offers a faster, cheaper, and simpler method for characterizing these more complex targets and aiding future development of novel therapeutic agents.
The Future of AlphaFold and Protein Prediction
The success of AlphaFold has understandably generated excitement in the scientific community. Although it enables researchers to pursue new protein targets, there are currently no new drugs that have been developed as a result of this technology. Therefore, it may still be too early to evaluate the full extent of AlphaFold’s impact on drug discovery until more time has passed. However, it has resulted in important changes for two ongoing challenges: characterizing the shapes of existing proteins and creating new ones from scratch. First, multiple comparable structural prediction models have emerged as a result of DeepMind’s innovation to further refine the industry’s capacity for accurate protein modeling. Second, AlphaFold has recently led to the creation of AI-generated diffusion models, such as DALL-E; these support the design of entirely novel proteins which are intended either to be drugs themselves or to help catalyze new drugs. Ultimately, these early advances in AI-powered protein modeling technology hold significant potential for inspiring creativity and innovation within the field of novel drug discovery and design.
As machine learning technology continues to develop, preclinical and potentially clinical phases of drug development will evolve simultaneously. Vial is a full-service CRO that recognizes the role of technology in the future of drug development and is paving the way for modernized clinical research through digital innovation. Trusted by leading sponsors, our specialized teams deliver shorter study timelines, quality affordable services, and a clinical trial experience that puts you first. Contact a team member today to discover how we can help!