Growing up, PolarisQB Principal Software Engineer Maurice Benson was a huge fan of Marvel, particularly the Iron Man series. In this hero, Benson saw a smart scientist in a lab who could achieve almost anything, provided of course he had almost infinite resources including the fastest computers in the world at his disposal. That is what started a lifelong passion, and led him to study machine learning (ML) and artificial intelligence (AI) and learn the ways that science could solve complicated problems with powerful computation. In his professional life, he has pursued this passion in many arenas, from fingerprinting cyber threats to training models to recognize optimal times to harvest crops. At PolarisQB he is applying his expertise in ML and AI to recognize traits in molecules and search massive chemical libraries with the most powerful tool he has ever had at his disposal, the Quantum Annealing computer.
When Bill Shipman, PolarisQB CTO, reached out to Maurice in 2020 with the idea of using Quantum Annealers to increase the speed of drug discovery for PolarisQB, he was intrigued. Chemists wanted calculations that would take months of cloud compute hours to complete, and time was most certainly of the essence. These calculations represent life-changing treatments and medicine for people suffering from myriad hard-to-treat diseases around the globe. Though quantum computing is just transitioning from its infancy as a strictly experimental technology to a commercially viable service, he saw promise in using these computers to turn months-long calculations into quantum calculations that could be solved in a matter of minutes. As Benson describes it, “Our chemists want to explore billions of molecules at a time. To brute force this problem on classical cloud computers it would take literally months costly and time-consuming compute hours to process. On the quantum computer, we get a sample of the best answers back in seconds!” This kind of acceleration in computational drug discovery processes represents a paradigm shift for the industry empowering chemists to explore a larger space in higher fidelity in less time.
PolarisQB has amplified our cloud computing platform Tachyon with the power of Quantum computing. The Tachyon system is able to take a reference molecule, calculate the various ways it could be processed, and start autonomous agents in the cloud to conduct the calculations in parallel. As PolarisQB chemists discover new ways to score our molecules, Tachyon is able to integrate this new information and run these calculations at scale. This hands-off automation allows PolarisQB to analyze vast chemical space (10^23 molecules) without needing an army of computational chemists. Benson characterized the current state of computational chemistry for drug discovery, saying, “In this current market good chemists are expensive but cloud computing is cheap. Leveraging our autonomous agents and quantum computing turns each one of our chemists into a superhero. With minimal input, our chemists will have a good idea of where to explore next in a matter of minutes instead of weeks or months”.
