Artificial intelligence has created a new drug in just 21 days

Artificial intelligence has succeeded already in many areas of our lives. And basically the AI finds its application in the field of complex calculations, mathematical models and so on. However, in the medical field and artificial intelligence may be equally useful (do you follow the progress of the AI? We regularly talk about them in our Telegram channel). For example, recently a Hong Kong company Insilico Medicine published the results of a study showing that its system based on AI and deep learning can create new drugs against certain abnormalities in just 3 weeks. And this is several tens of times faster than traditional methods.

How artificial intelligence creates medications

AI from the company Insilico Medicine is called GENTRL. And remarkably, at the helm of the company is our compatriot Alex Zhavoronkov. Mr. larks in the mid-2000s, received a master’s degree from Johns Hopkins University and then his doctorate at Moscow State University where his research was focused on using machine learning to study the physics of molecular interactions in biological systems. In 2006, he has worked at ATI until it was acquired by chip maker AMD. In 2014, Alex co-founded the already mentioned Insilico Medicine, having experience in the technology industry and interested in the issues of pharmacy.

If you go back to AI, then the developers are called the core technology of artificial intelligence “generative tensor learning”. It allows the AI, if you do not go into the details more efficiently and quickly learn the required skills.

We thought: can we make the machines to invent from scratch new molecules with certain properties instead of forcing them to sort through dozens of available options, says Alex Zhavoronkov.

Insilico used GENTRL in order to create multiple (or to be exact, 6) options of drugs for the treatment of muscle fibrosis. Created drugs inhibit DDR1 receptor, which is involved in the development of the disease. For this AI took 21 days, after which the scientists chose the most suitable versions of the drugs and tested them on laboratory animals. It took another 25 days. Thus, the selection of potential drugs, it took just 46 days. For comparison, the traditional development process of candidates for the title of the medicine takes about 8 years and costs several million dollars. While the AI took only 150 thousand dollars.

Left — normal muscle tissue. To the right is the fabric with the development of fibrosis