Accelerating Efficient Drug Discovery With the Power of AI

The rise of artificial intelligence (AI) is beginning to revolutionize many industries, from IT to marketing – as well as drug discovery.

Now, a recent study has showcased the capabilities of AI in the biotechnology and pharmaceutical industries, accelerating the design and production of a drug designed to treat hepatocellular carcinoma (HCC), a type of liver cancer. The researchers focused on cyclin-dependent kinase 20 (CDK20), a protein they found was associated with HCC but that also has no established 3D structure. Using the computer program AlphaFold combined with other AI-based approaches, they were able to predict the protein’s 3D structure and generate molecules to target it.

To find out more about this approach, and the promise that generative AI holds for the drug discovery industry, we spoke to Dr. Alex Zhavoronkov, senior author of the study and founder and CEO of Insilico Medicine.

Sarah Whelan (SW): Could you briefly explain to our readers that may be unfamiliar, what is AlphaFold and what can it be used for?

Alex Zhavoronkov (AZ): AlphaFold is a computer program from Alphabet’s DeepMind that has predicted the protein structure of every known protein using AI. This free AI-powered database can predict the 3D structures of millions of proteins from their primary amino acid sequences with an accuracy that rivals experimental methods. AlphaFold – and now AlphaFold2 – is considered a major breakthrough in AI and structure-based biology which we think can help accelerate the development of new medicines. When we first learned of this breakthrough, we were eager to apply it to our own AI platform to see if we could further accelerate our target discovery and novel drug design, particularly for a target without a known experimental structure. 

SW: Could you tell us more about how this technology was applied to discover a novel HCC target and hit molecule? Why is CDK20 an attractive target?

AZ: As noted in the paper, there are limited CDK20 inhibitors reported despite great success being achieved with inhibitors against other members of the CDK family. One possible reason is that there is no available 3D structure for this target. By combining the AlphaFold predicted protein structure for CDK20 with our generative AI drug design platform, Chemistry42, we identified possible binding sites for a small molecule inhibitor of CDK20. Ultimately, we wanted this research to demonstrate that it is possible to use a predicted structure for a novel target and come up with usable chemical data. We not only accomplished that goal, but we did it in less than 30 days. 

SW: What are the advantages of this approach over other drug discovery methods?

AZ: AI – and particularly AI coupled with robotics – introduces speed and efficiency into the drug discovery and design process, while also helping to control costs. In traditional drug discovery, it takes over 10 years and costs around $2 billion to bring one drug to market – and 90% of drug candidates fail during human trials. This high cost and slow speed is preventing new life-saving medications from reaching patients.  

Insilico Medicine utilizes generative AI to understand the mechanisms behind many different diseases to discover new molecular targets and pathways. Our platform then “imagines” new molecules that don’t exist in nature and can become the next breakthrough drugs. It’s like combining ChatGPT for biology, which can work with text and biological data, and Midjourney for chemistry, which can create the molecules with the desired properties. We began these efforts in 2014 and invested deeply and early into AI and have accumulated a lot of data. We followed roughly $2 trillion worth of research data and invested significant time and resources in making the data machine-learnable so that it can be used in our AI platform. We then released several software AI tools that are now used by many academics, pharmaceuticals and biotechnology companies – as the platform is used it is continually validated and improved, in a constant positive feedback loop. 

SW: What are the main challenges and limitations of using these kinds of AI approaches?

AZ: Perhaps the major limitation is remaining skepticism in the industry, although AI is increasingly being explored by most big pharmaceutical companies eager to advance their own programs and accelerate the pace of drug discovery. There is still no AI-generated small molecule drug treating patients, which many are waiting for.

I should note that Insilico Medicine is the first company with a drug discovered and designed by generative AI. It has gone through many steps of experimental validation and is now nearing Phase II trials in human patients, which is very exciting for us and for the industry. This lead drug candidate, for the devastating lung condition idiopathic pulmonary fibrosis, just received Orphan Drug Designation from the FDA.

SW: What do you think the future of AI in drug discovery looks like? What do you think can be achieved with this technology?

AZ: I believe the adoption of AI by pharmaceutical companies will continue, as will partnerships between pharma companies and biotechs like Insilico. Increasingly, I believe it will be important to provide pharmaceutical companies with full end-to-end capability, from target identification to drug design to clinical trial prediction in order to create one seamless pipeline. I also believe robotics will increasingly work in concert with AI to bring additional speed and capacity to drug development. Our own AI-run Intelligent Robotics Lab was just opened and will help to further accelerate our efforts. Finally, I believe we are fast approaching the era of truly personalized medicine, where treatments can be designed for a person based on his or her individual health and genetic profile.

Dr. Alex Zhavoronkov was speaking to Dr. Sarah Whelan, Science Writer for Technology Networks.