Deep Quantum Modeling: Q&A with Adityo Prakash

Adityo Prakash
CEO
Verseon

While AI might be taking up all of the headlines, there are other advanced technologies helping to improve the drug discovery process. Verseon is combining AI with deep quantum modeling, which allows for stronger models of protein-drug interactions. CEO Adityo Prakash spoke with Pharmaceutical Executive about how his company is utilizing this new technology.

Pharmaceutical Executive: How does Deep Quantum Modeling work?
Adityo Prakash: The objective of drug discovery is to find molecules that bind to a disease-causing protein target and change its function. Today, the overwhelming majority of small-molecule drugs are found via High-Throughput Screening, a trial-and-error process that tests the target protein against a range of previously synthesized drug-like molecules.

The set of distinct drug-like chemotypes that the pharma industry has made and tested over the past 150 years is between 7 and 10 million. But this number pales in comparison to the entire set of one decillion (a one followed by 33 zeroes) drug-like small molecules possible under the rules of organic chemistry. And since AI can only predict molecules similar to those in its training dataset, it cannot help the industry find fundamentally new small-molecule drugs.

Over the past 30 years, the stretch goal of the pharma industry has been to design new drugs ab initio by constructing completely novel molecules on the computer that selectively bind to the target protein. However, the tools to perform this task with sufficient accuracy have not been available.

The physical interactions that govern protein-drug binding in an aqueous environment are governed by quantum mechanics. It is a fiendishly complex problem that is not computationally tractable using a brute-force approach. Fundamental new insights into molecular physics are necessary in order to simulate these interactions with sufficient accuracy while keeping the models computationally tractable. Verseon’s Deep Quantum Modeling™ (DQM™) is based on groundbreaking advancements in molecular physics to calculate with stunning accuracy all necessary aspects of the interactions between a drug-like molecule and a target protein.

DQM™ opens the door to finding entirely novel drug-like molecules without the need for prior experiments or training data.

PE: What is this technology’s relationship with AI?
Prakash: Deep Quantum Modeling™ drives innovation, while AI performs optimization.

Unlike AI, DQM™ does not require prior experimental training data to explore billions of novel drug-like molecules on the computer. DQM™ is the starting point of our discovery process.

We synthesize the most promising molecules found through DQM™ and test them in the lab. These tests yield data on entirely novel classes of drug-like molecules that have never before been made or tested elsewhere. We in turn train our AI system VersAI™ on this new data.

VersAI™ can then do what AI does best: find variants of our novel compound families during the optimization process. The best variants validated in the lab are then advanced into clinical trials.

PE: What are some of Verseon’s success stories?
Prakash: All the candidates in our pipeline were discovered and developed using DQM™ and VersAI™. Each candidate in our expanding pipeline offers a uniquely beneficial therapeutic profile, with the potential to redefine the standard of care for the disease it addresses.

Take our Precision Oral Anticoagulants (PROACs™), for example. All anticoagulants on the market today pose high risks of uncontrolled bleeding events, making them unsuitable for long-term coadministration with antiplatelet drugs. But our drug candidates preserve normal platelet function and near-normal bleeding in response to injury while preventing the internal clots that cause strokes and heart attacks. These drug candidates promise to make long-term administration of blood thinners safe for the world’s 400 million cardiovascular disease patients, whether administered alone or with an antiplatelet drug.

Our drug candidates for diabetic retinopathy (DR) would be another interesting example. Currently, one third of all diabetics experiencing progressive vision loss have to wait until disease is severe enough to justify the risks of regularly injecting repurposed cancer drugs into their eyes–risks like corneal melting, geographic atrophy, cataracts, inflammation, infection, and increased intraocular pressure. And these drugs only treat the symptoms–not the root cause–and do not work for 50% of patients. Using our DQM+VersAI platform, we’ve developed the first and only drug candidates shown to prevent and reverse diabetic retinopathy. These small-molecule drug candidates can be administered orally as prophylactics at the earliest signs of vision loss, preventing and reversing disease progression in 179 million DR patients.

Our drugs for anticoagulation and diabetic vision loss are just two examples of what our platform can do. We’re generating a steady stream of life-changing new medicines for a range of major human diseases.

PE: How important is it for Pharma companies to experiment with new technologies and experiment with combining new technologies together?
Prakash: While there has been a lot of progress in our understanding of biology and disease processes, the pharmaceutical industry has so far managed to develop treatments for only 500 out of roughly 10,000 known human diseases. And these treatments all too often work poorly or have a long list of undesirable side effects.

Needless to say, medicine has a long way to go in giving humanity the much healthier future it deserves. Accelerating the development of new treatments requires the integration of multiple new technologies–not just using a single tool alone. Complex problems like drug discovery and development require significant advances in a range of scientific and technological disciplines. Companies that manage to make innovations in areas like physics, engineering, chemistry, biology, and AI work in concert will lead the future of pharmaceutical medicine. And Verseon plans to be foremost among them.