How Insilico Medicine uses AI to accelerate drug development

Within the longevity research community, Alex Zhavoronkov is well known for his relentless focus. He works seven days a week and does not take vacations. The hard work pays off: In February, Insilico Medicine, the AI ​​drug development company he founded, announced the first phase 1 clinical trials for a drug developed entirely by AI. After a series of investment rounds throughout the rest of the year, the company is now well funded and its software is widely used in the pharmaceutical industry. Alex explains the company’s progress in the latest episode of the London Futurist Podcast.

Three phases of drug development

Drug development is a long and complicated business, but you can break it down into three phases. First, you create a hypothesis about what causes a disease. The culprit is usually a protein, or a set of them. There are over 60,000 targets to hit and 10,000 diseases. This first phase is the most important of the three, and is usually done in academia, where it costs billions of dollars to develop a hypothesis, and 95% of them turn out to be wrong. Insilico created a software platform called Pandaomics to generate these hypotheses. It includes more than twenty AI models, which are trained on huge public data repositories.

The second phase is to develop a molecule that could treat the disease. Insilico’s platform for this second phase is called Chemistry 42. Where traditional pharmaceutical companies will test thousands of molecules to see if they bind to the target protein, AI allows you to watch how a protein folds up and imagine a molecule that could bind to it. . It is then enough to test a dozen molecules with the desired characteristics. Insilico does this with a variety of Generative Adversarial AI Networks, or GANs, as well as some Transformer models. Alex describes his collection of 40 validated models as a “zoo”.

The third phase of drug development is testing for efficacy and safety. Insilico’s platform for this is called InClinico, and it reuses some of the AI ​​models from the other two platforms, although it’s actually the company’s oldest platform. . It is trained on massive datasets on past clinical trials and has been validated on drugs that have made it all the way through the pipeline.

Massive savings of time and money

Insilico and other companies using AI to develop drugs have reduced the time and cost of the process by 90%. More importantly, they increased the success rate of drug candidates. So why hasn’t the entire pharmaceutical industry adopted his processes? Alex says the big pharmas all have AI groups and they’re trying to change, but they’re huge organizations and it takes time.

He gives the example of a major pharmaceutical company that validated InClinico for several months, but decided not to use it anymore because it could not bring itself to let outsiders determine the fate of its multi-billion development programs. There are no such worries, however, in hedge funds and banks betting on biotech startups, so for now, this is the sweet spot for Insilico.

One way to shorten the drug development timeline is to reallocate existing drugs to new diseases because they are already known to be safe in humans. The problem with that is that either someone else owns the intellectual property (IP) of these drugs, and you have to license them, or they’re already generic, which means there’s little chance of winning a commercial return on development costs.

Overall it’s good

Insilico is first and foremost a global organization. Alex was born in Latvia, he studied in Canada, he started his career in the United States and he founded Insilico in Hong Kong. Using Contract Research Organizations (CROs) in China has allowed Insilico to do research without having its own wet lab. It’s also easier in some ways to do clinical studies in China, although the company has to duplicate them in other territories that may not accept foreign data. Alex reports that Hong Kong still enjoys excellent intellectual property protection, deep financial expertise, world-class scientific resources and is a beautiful place to live. Post-Covid, Insilico has also set up a site in Shanghai.

Alex is frustrated by the growing resistance to international cooperation in pharmaceutical research. Great research projects need all the money and talent in the world, and when they succeed, they help the whole world. For Alex, the most important of these large-scale research projects is to understand and prevent aging. It is vital, he argues, to make clinicians aware of longevity research, and to that end he and a few colleagues have developed the Longevity Medicine Course. Alex’s relentless focus is underpinned by his belief that longevity research is the world’s most valuable philanthropic activity.