Don Van Dyke on the power of AI to transform drug discovery efficiency

How Cloud Pharma helped GSK drop time from target to lead molecule from four years to just four months

Pharma IQ

In recent years, it has been hotly debated whether artificial intelligence really has the power to transform the drug discovery space. Cloud Pharmaceuticals believes that not only is there potential for this change, but that they are already seeing the impact first-hand. 

Having used their own technology to partner with companies to design, develop and license novel compounds, they believe that the industry may be able to reverse the plummeting levels of drug productivity. 

In this interview, Don Van Dyke, Chief Operating Officer at Cloud Pharmaceuticals, tells us about the driving force behind these types of development, why some companies are still skeptical and what will be the turning point for AI adoption. 


Pharma IQ: Can you tell us a little bit about Cloud Pharma and what you’re doing in the AI space?

Don: “So, Cloud Pharmaceuticals designs novel drug molecules using various types of artificial intelligence.

The core of this was invented at Duke University by our founder. What’s unique about our approach is the way we are able to access novel molecular space that machine learning and other AI are not able to do.

This has led us to have several successes in designing across a broad variety of drug targets and led to commercial success, with partnerships with companies such as GSK.”


Pharma IQ: Why is there a need to use AI during the drug discovery process?

Don: “Well, drug productivity has been declining since 2000.

It’s interesting because when we engaged with GSK they told us that the process of going from target to lead molecule traditionally takes them four to five years. But using our process and technology, we were able to reduce that down to a matter of months. That also greatly reduces the dollar and time cost as well.

Using a targeted drug design logic and process also allows us to have a greater probability that the molecules we design will progress through the drug development process, rather than the uncertainty of traditional methods.”


Pharma IQ: Why do you think that some companies are shying away from the use of AI?

Don: “Well, it’s a new thing. I’ve been involved in the scientific research market for most of my career and I think that scientists and drug companies are always careful to choose methods which have been validated. So you get a bit of a chicken and egg problem.

We tried to solve this through our work with universities where they were able to validate our technology.

It’s a sensible caution that the companies have and that gets compounded by many new companies making claims that can be a little exaggeration or incorrect. This only adds to the confusion about how to choose the right technology or company to partner with.

Although I do think this is normal and we are going through a sorting out phase now.”


Pharma IQ: Do you think as more successes are shown, that the industry will come around to the idea?

Don: “Yes. Exactly. When you show proof, then that’s the way the market goes.

It’s very much word of mouth, about who has succeeded from what technology – when you see demonstrated results, then embracing technology is much easier.”


Pharma IQ: As an industry, how far away are we from that point?

Don: “I think we’re right at that point now. We’re able to demonstrate going from target to lead molecule  on several different programs, so I think we’re very close.”


Pharma IQ: What progress has Cloud Pharma been able to make so far?

Don: “Well, we did 15 different programs at the University of Florida on a vast variety of target types and we’ve been able to do some really cool things. We’ve been able to hit a previously un-druggable target, the Beta Common Receptor and that happened to be a protein-protein interaction that we were able to successfully disrupt.

We’ve been able to design a selective PDHK2 inhibitor. We’ve done some things on novel targets and we’ve done some things on selectivity and protein-protein interactions that seem to be gaining a lot of interest.”


Pharma IQ: Do you think AI will be able to move past the traditional barriers faced in the drug discovery process?

Don: “Absolutely. You have two dynamic issues.

One is that the druggable genome is at around 3,000 and the number of targets that have been drugged is only in the hundreds. So we have an order magnitude gulf between what has been drugged and what hasn’t, but would be beneficial to be drugged.

On the other side, we have vast molecular space where somewhere out in the virtual universe, molecules exist virtually that could assist in drugging these previously un-druggable targets.

The task is to sort through this vast molecular space to find these novel molecules that will do this.”


Pharma IQ: What do you think the industry is going to be most focused on next, in regards to AI?

Don: “I think what’s going to happen is that we’re going to see the prosecution of a lot more targets in a shorter period of time. Which will lead to a change in the sort of metrics that the pharma companies, start-ups and biotechs use, when they now have an agile, fast and inexpensive way to go after their targets.

It will change the approach that companies take.

I think they’re going to be less risk averse in prosecuting new targets because they’ll be able to do it quickly and less expensively. And that’s what we all want; the ability to find good molecules and to cure patients.“