THANK YOU FOR SUBSCRIBING
Pharma Tech Outlook | Tuesday, May 17, 2022
Novartis is working with Microsoft to apply machine learning to medicinal chemistry as part of an effort to use AI to get treatments to patients faster.
FREMONT, CA: A common misconception about drug hunting can obscure one of the most difficult challenges chemists face when developing new therapies. Drug hunters must typically create new drugs from scratch to create first-of-their-kind medicines for diseases with no treatments. It's a time-consuming process that involves creating and testing thousands of experimental compounds before finally finding one that's safe enough to be tested on humans.
Stay ahead of the industry with exclusive feature stories on the top companies, expert insights and the latest news delivered straight to your inbox. Subscribe today.
Chemists must design and synthesize molecules with specific properties in mind during the early stages of drug discovery. For example, they must be effective against a specific biological target (usually a key protein suspected of contributing to the disease) and be soluble and well-tolerated in the human body. Every new set of potential medicines is tested in a series of experiments to assess these attributes. Scientists use the resulting data to determine which synthesized molecules are the most promising and then use what they've learned to design the next round of new, better-optimized compounds. This time-consuming "design, make, test, analyze" cycle can take years, but there's reason to believe AI – artificial intelligence – can help.
Two years ago, Novartis began working with Microsoft, a leader in machine learning, to leverage game-changing digital technologies to help deliver medicines to patients faster. It's a collaboration that has resulted in a digitally-charged "prototype" Generative Chemistry pipeline that has already been implemented in a diverse set of medicinal chemistry projects, some of which are bearing fruit.
Those involved in the collaboration emphasize that machine learning will not replace scientists' expertise, experience, and intuition. Instead, AI is expected to supplement human knowledge. The goal is to reduce the time and effort required to find and analyze relevant data and facilitate a give-and-take between human researchers and AI.
Furthermore, those behind the collaboration anticipate that machine learning will aid in predicting which formulation designs have the best chance of being effective and reveal which experimental parameters are most useful in specific drug development scenarios. However, to maximize the benefits of AI in drug discovery, it must be made available across all research programs. To that end, Novartis has collaborated with Microsoft to make the tools available to researchers with limited computer experience.
More in News