In an unprecedented advancement in drug discovery, Zapata Computing, Inc., alongside Insilico Medicine, the University of Toronto, and St. Jude Kid’s Research Hospital, has showcased the remarkable potential of quantum-enhanced generative AI. This collaboration has led to the first-ever instance where a generative model operating on quantum hardware surpasses traditional classical models in generating viable cancer drug candidates.
This landmark study focused on developing novel KRAS inhibitors, a notoriously difficult goal in cancer therapy. Utilizing advanced generative AI models on each classical and quantum hardware, including a 16-qubit IBM device, the team successfully generated a million drug candidates. Following a meticulous technique of algorithmic and human filtering, the quantum-enhanced generative model yielded two distinct molecules with superior binding affinity over those produced by classical models. This breakthrough not only underlines the efficacy of quantum computing in drug discovery but in addition illustrates the transformative role of Industrial Generative AI in addressing complex, domain-specific challenges in various industries.
Industrial Generative AI, a specialized subcategory of generative AI, is especially adept at tackling such intricate problems. Unlike general-purpose AI tools like ChatGPT and DALL-E from OpenAI, Industrial Generative AI is customized to handle specific issues inside enterprises or industries. It navigates through challenges corresponding to data disarray, large solution spaces, unpredictability, time sensitivity, compute constraints, and demands for accuracy, reliability, and security. At its core are generative models, like Large Language Models (LLMs), which learn from training data to generate recent, realistic outputs. This approach is what enabled the Zapata AI team to pioneer in the sector of drug discovery, leveraging AI to create groundbreaking solutions.
Yudong Cao, CTO and co-founder of Zapata AI, highlighted the synergy of quantum and classical computing in providing comprehensive solutions on this groundbreaking project. The research, currently awaiting peer review and available on ArXiv, builds on earlier studies demonstrating the potential of quantum generative AI in drug discovery.
Alex Zhavoronkov, PhD, founder and co-CEO of Insilico Medicine, acknowledged the mixing of Insilico’s generative AI engine, Chemistry42, with quantum-augmented models, heralding recent therapeutic avenues for difficult cancer targets. This step is critical in advancing the longer term of drug discovery.
With a recent strategic partnership with D-Wave Quantum Inc., Zapata AI is about to further expand the horizons of quantum generative AI models in discovering recent molecules for a variety of economic applications. Christopher Savoie, CEO and co-founder of Zapata AI, expressed excitement about this development and the potential for broader application in various industries.
Alán Aspuru-Guzik, a professor on the University of Toronto and a co-founder and Scientific Advisor of Zapata AI, shared his optimism about integrating quantum computing into the drug discovery pipeline. This research is pioneering, setting a precedent for future quantum computers to showcase their unique capabilities.
The research employed Zapata AI’s QML Suite Python Package, available on its Orquestra® platform, emphasizing the sensible application of quantum computing in solving real-world scientific challenges. This integration of Industrial Generative AI into the drug discovery process marks a major stride in leveraging AI for modern, industry-specific solutions, driving growth and efficiency within the ever-evolving technological landscape.
