
Introduction
The pace of discovering treatments for human diseases is slow, meaning countless medical needs remain unaddressed. Meeting this challenge calls for a bold and fundamentally new way to identify medications capable of combatting both acute and chronic illnesses.
Here, Adityo Prakash, CEO of pioneering drug discovery platform, Verseon, shares his perspective on the challenges facing the pharmaceutical industry and their innovative solution to drug discovery:
The Verseon platform successfully integrates physics and maths with proprietary lab data and AI. Can you explain how this integration produces new drug candidates?
Many of the more widespread methods and trial and error experiments we see in use are simply too slow, so we developed a more novel approach to discovering and producing new drug candidates.
Over the course of two decades, we developed significant advances in multiple distinct areas of science to build our molecule engineering platform. On the computer, we design billions of novel drug-like molecules that have never actually been made before. Using our breakthroughs in quantum physics-based molecular modelling, we predict which of these new molecules will bind to a target protein. Our physics modelling identifies completely novel drug molecules that we make and test in the lab.
We then use the chemical and biological data from these new molecules to train our own AI. Our AI generates variants of these novel molecules, which we further validate in the lab. We choose the best of these to advance to clinical trials.
For every program we undertake, this process generates multiple clinical candidates with unique therapeutic profiles. These are candidates that cannot be found through any other current method.
Many companies use AI for drug discovery, yet success rates remain extremely low. How does the Verseon approach differ and how or why is your model more successful?
As many experts have pointed out, AI is good at interpolation and terrible at extrapolation. AI trained on the existing limited experimental data can only generate molecules that are highly similar to those in its training dataset.
As a result, current AI efforts are making small tweaks on existing drugs or repurposing old drugs for new indications. But AI’s ability to create entirely new drug molecules is fundamentally limited.
The industry needs to abandon the idea that using AI alone can lead to truly novel drugs – instead, physics-based modelling needs to be the first step in drug discovery. Unlike AI, physics-based modelling, if done right, would not require preexisting experimental data to generate drug-like molecules that belong in previously uncharted regions of the chemical space. Results from the physics-based design of new molecules can then be used to train AI to create further variants.
The concept of physics-based molecular modelling is not new. Other organizations like Schrӧdinger and CCDC have been selling physics-based modelling tools for almost three decades. However, existing tools lack sufficient accuracy to make them the primary driver of drug discovery. Verseon is the only company to break the accuracy barrier and make systematic drug discovery a reality. And unlike AI-first or AI-only efforts, Verseon’s approach does not require any prior experimental training data to produce promising new drug candidates.
Verseon systematically designs molecules before testing them. What proportion of designed compounds typically progress to a meaningful outcome?
In most cases of drug development, only about 15 per cent of lead compounds advance into clinical trials, and only around 10 per cent of those become medicines. It’s a very long process and decisions have to be made at every stage. It’s vital to get it right, or compounds with good potential could end up going to waste while unnecessary time, effort and resource is spent on those with no promise.
Verseon now has 16 drug candidates in development. Which candidates are closest to launch and in what timescale?
We expect two of our programs to complete clinical trials and reach market around the same time. The first program is developing what we call Precision Oral Anticoagulants, or PROACs, to prevent heart attacks and strokes. Our new class of compounds prevents the formation of hazardous blood clots responsible for strokes and heart attacks.
The second program addresses diabetic vision loss, which affects one in three diabetics. Diabetes causes leaks in retinal blood vessels that damage the retina and eventually lead to vision loss. The current standard of care is to wait until the patient is almost blind to justify the risks of regular injections with a repurposed cancer drug—into the eye. Instead, our diabetic vision loss drug candidates can be delivered as oral pills.
Several of your candidates target conditions with exceptionally large patient populations, such as cardiovascular disease and diabetic complications. What clinical advantages do you expect these drugs to offer compared with current medications on the market?
Firstly, the PROACs we’re developing to prevent heart attacks and strokes do not increase the risk of bleeding, unlike currently available anticoagulants. This unique property makes them suitable for safe lifelong use, either alone or in conjunction with antiplatelet drugs. PROACs have the potential to change the standard of care for a sizable portion of the world’s 23 million cardiovascular disease patients.
Secondly, our diabetic vision loss drug candidates will be a gamechanger for patients. The oral pills mean no injections, and no waiting for disease progression. Half of all patients receiving the currently used form of treatment experience no improvement in their symptoms, but published test results show that our drug candidates reverse retinal vascular permeability in vivo. When our oral drugs reach market, they will change the standard of care for the world’s 196 million diabetic retinopathy patients.
Just those 2 programs have the potential to improve the lives of over half a billion people worldwide. And as our drug programs for other indications like cancer enter the market, we will touch the lives of many more people.
What impact could systematic drug discovery have on health systems already facing rising cost pressures, like the NHS?
We know the NHS is under huge cost pressures, and we’re hopeful that systematic drug discovery will have a real impact in improving outcomes for patients.
Systematic drug discovery essentially streamlines the process, in theory leading to more efficient discoveries and better use of resources. It helps to optimise the whole process, which should mean fewer failures and more successful treatments, ultimately reducing the burden on the NHS and helping limited funds go further.
The current small-molecule drug discovery that is often used is primarily reliant on trial-and-error experiments in the form of high-throughput screening. This approach has produced some important treatments, but overall it is slow, costly and fraught with uncertainty.
Anything we want to test as a potential drug candidate has to be made in the lab first, and making new molecules that haven’t been synthesised before is an expensive process. The future of great medicines lies in the unexplored chemical space, and we’re hopeful that it will be a truly cost effective solution to improve the lives of patients across the globe.