Artificial Intelligence (AI) is reshaping operations across various industries, including the pharmaceutical arena. Although AI-enabled drug discovery is still in its infancy, ongoing investor interest and technological advances suggest an increasingly pivotal role in developing new therapies in the future.
AI-Enabled Drug Design Methods
Traditionally, drug discovery involves a time-consuming and costly process of target identification, compound screening, and preclinical testing. The process can take an average of five years to complete. AI-assisted drug discovery methods have the potential to significantly cut time and expense.
Unlike the human workforce, machine learning algorithms can analyze vast datasets in seconds, identifying patterns and predicting how different compounds interact with biological targets. Deep learning models, particularly neural networks, can learn from existing data to generate new molecular structures.
Natural language processing tools can sift through reams of clinical trial data and scientific literature to uncover insights. Together, these technologies are already showing promise.
First AI-Assisted Drugs to Enter Clinical Trials
UK-based Exscientia uses an AI-driven drug discovery platform to develop cancer and neurodegenerative disorders treatments. Exscientia and Japanese firm Sumitomo Dainippon Pharma designed DSP-1181, known as the first AI-assisted drug to enter clinical trials.
Designed to treat obsessive-compulsive disorder, the full serotonin 5-HT1a receptor agonist was able to move into clinical testing within just one year—as opposed to the average five years it takes using traditional drug discovery methods.
AI-driven technology also aided the discovery of a small-molecule inhibitor of TNIK. Studies have found that the drug ameliorates fibrotic processes in kidney, lung, and skin fibrosis disease models and improves lung function in a murine model of lung fibrosis. Biotech firm Insilico Medicine designed INS018_055 using the commercially available target-discovery platform, PandaOmics. In just 18 months, the drug went from target discovery to preclinical candidate.
$850M Recursion Exscientia Merger
In early 2024, Recursion Pharmaceuticals introduced BioHive-2, a supercomputer powered by Nvidia technology designed to expedite drug discovery. On August 8, Recursion announced it had bought Exscientia for $688 million.
The potential acquisition would provide access to partnerships with Sanofi, Merck KGaA, GSK, and Bristol Myers Squibb and leave the combined entity with approximately $850 million to support three years of operations.
Challenges with AI-Driven Drug Design
While progress has been significant, AI's integration with drug discovery faces several challenges.
The quality and quantity of data available for training AI models are not there yet. Drug discovery relies on high-quality, diverse datasets, and integrating data from various sources remains complex.
Additionally, AI models are often seen as "black boxes." Their decision-making processes are not always transparent. This lack of interpretability can disrupt our understanding of how AI models arrive at specific conclusions, which is crucial for regulatory approval, intellectual property concerns, and scientific validation.
Still, the potential benefits of AI in drug discovery cannot be denied. The ability to analyze large datasets, predict compound interactions, and streamline the discovery process could bring more efficient and cost-effective drug development. As AI technologies evolve, their integration into drug discovery and design should become more refined and impactful.