top of page
  • AND

Using generative AI and quantum computing to speed up drug discovery

Insilico Medicine, a clinical-stage company specializing in generative AI-driven drug discovery, recently published a study in the Journal of Chemical Information and Modeling. The study, led by Insilico's centers in Taiwan and the UAE, aimed to explore the potential benefits of combining quantum computing and generative AI in lead candidate discovery for drug development. The research, supported by Alán Aspuru-Guzik, PhD, from the University of Toronto Acceleration Consortium and scientists from the Hon Hai (Foxconn) Research Institute, showcased the successful application of quantum generative adversarial networks in generative chemistry. Insilico Medicine focuses on pioneering groundbreaking methods and engines utilizing rapidly advancing technologies, such as generative AI and quantum computing, to accelerate the process of drug discovery and development.

''This international collaboration was a very fun project. It sets the stage for further developments in AI as it meets drug discovery. This is a global collaboration where Foxconn, Insilico, Zapata Computing, and University of Toronto are working together."

Generative Adversarial Networks (GANs) have emerged as highly effective models in drug discovery and design, exhibiting impressive capabilities in generating data that closely resembles a given data distribution across various tasks. The traditional GAN architecture comprises a generator and a discriminator. The generator takes random noise as input and aims to replicate the underlying data distribution, while the discriminator endeavors to differentiate between real and fake samples. Training a GAN continues until the discriminator can no longer distinguish between the generated and real data.

In this research paper, scientists investigated the potential quantum advantage in small molecule drug discovery by progressively substituting each component of MolGAN, an implicit GAN for small molecular graphs, with a variational quantum circuit (VQC). This process involved replacing the noise generator, the generator employing the patch method, and the quantum discriminator, and comparing their performance with their classical counterparts.

The study not only demonstrated that trained quantum GANs can generate molecules resembling those in the training set by employing the VQC as the noise generator, but also revealed that the quantum generator surpasses the classical GAN in terms of the drug properties of the generated compounds and the achievement of specific benchmarks. Furthermore, the study revealed that the quantum discriminator of the GAN, which possesses merely tens of learnable parameters, can generate valid molecules and outperforms the classical counterpart, which requires tens of thousands of parameters, in terms of both the properties of the generated molecules and the KL-divergence score.

''Quantum computing is recognized as the next technology breakthrough which will make a great impact, and the pharmaceutical industry is believed to be among the first wave of industries benefiting from the advancement. This paper demonstrates Insilico's first footprint in quantum computing with AI in molecular generation, underscoring our vision in the field."

Based on these discoveries, scientists at Insilico have outlined plans to incorporate the hybrid quantum GAN model into Chemistry42, the company's exclusive engine for generating small molecules. This integration aims to further enhance and expedite Insilico's AI-driven drug discovery and development process.

Insilico Medicine was among the pioneers in utilizing GANs for de novo molecular design and published the initial paper in this field back in 2016. Leveraging generative AI models based on GANs, the company has successfully produced 11 preclinical candidates, with its lead program even undergoing validation in Phase I clinical trials.

"I am immensely proud of the positive outcomes achieved by our quantum computing team, thanks to their dedication and innovative approach," expressed Dr. Alex Zhavoronkov, founder and CEO of Insilico Medicine. "I believe this represents just the beginning of our journey. We are currently engaged in a groundbreaking experiment involving an actual quantum computer for chemistry, and we eagerly anticipate sharing Insilico's invaluable expertise with both industry and academia."


bottom of page