When artificial intelligence is used in the manufacture of medicines

Someone would think that AI is now all the rage. But in the pharmaceutical sector, change is more than just cosmetic. (Image: 123RF)

Paris – Looking for a cure for dengue fever using artificial intelligence (AI)? It’s not science fiction, but the project recently launched by a European NGO, a new sign that AI now has a place in medicine.

The NGO Drugs for Neglected Diseases Initiative, which seeks cures for neglected diseases, launched a partnership in April with BenevolentAI, a British company working to develop new molecules using AI. BenevolentAI was absent in its first attempt. Notably he revealed, during the pandemic, the role that a molecule, baricitinib, developed by Eli Lilly’s laboratory for another disease, could play in the treatment of COVID-19 patients.

Someone would think that AI is now all the rage. But in the pharmaceutical sector, change is more than just cosmetic. In early 2020, Exscientia, a Scottish start-up, developed together with Japanese pharmaceutical laboratory Sumitomo Dainippon which was the first molecule “built” by AI, entered clinical trials.

“It’s not futuristic: artificial intelligence is a methodological approach to data processing, which can be used at some stage of the drug industry’s development process,” Drs. Thomas Borel, director of scientific affairs. of the federation of pharmaceutical companies (Leem).

A visit to the Paris area of ​​French start-up Iktos, founded in 2016, alerts us to the changing weather. Here, there are no microscopes or biology devices, no lab technicians in white coats. But computer screens are so numerous, that they traverse a lot of health data at speeds beyond the reach of the human brain.

“The idea is to exploit existing data to get new interesting molecules, faster,” explains Yann Gaston-Mathé, the head of the start-up, which he founded in 2016.

For this, his team used a global database containing data from 100 million molecules. From here, “we trained a model that would automatically generate new molecules, and identify those that would be active in biological targets of interest,” Yann Gaston-Mathé describes.

Iktos has even set up a platform for researching molecules using artificial intelligence, which it offers on subscription to pharmaceutical companies.

Interest of large labs

Aqemia, a start-up from the École nationale supérieure-PSL, created in 2019, is developing a drug discovery platform using quantum-inspired statistical physics.

“We use an artificial intelligence that is said to be generative,” underlines founder, researcher Maximilien Levesque. “We invented molecules that stick to a specific biological target responsible for a disease. Artificial intelligence is fed by physics: we just need to know the physical properties of the molecule and the target to calculate the their connection, “he explained.

If start-ups are ahead, major laboratories are increasingly looking at the issue, and paying the price. American giant Bristol-Myers Squibb signed a deal with Exscientia last year, for which it could pay more than a billion dollars. GAFAMs are also involved: in 2019, Swiss laboratory Novartis and giant Microsoft announced their collaboration on the subject.

Is this for everyone the end of the chemist in his laboratory? There are huge challenges, including access to actionable data. Not forgetting the need to seek out data specialists in the future, experts in both artificial intelligence and pharmacology issues.

There is also an important aspect of regulation, judge Thomas Borel, of Leem. “For example, we use AI to create a virtual group of patients during a clinical trial. But in order to receive this drug, regulatory systems must recognize the value of the algorithm,” he said.

“Drugs have been designed with the help of computers for years,” comments Yann Gaston-Mathé, who says he wants to provide “additional tools to chemists, without wanting to replace the human machine”.

Leave a Comment