Recognizes side effects

Pioneering AI for pharmaceutical companies comes from Graz

Artificial intelligence is playing an increasingly important role in laboratories (symbolic image).

(Bild: stock.adobe.com)

Every year, almost 90 percent of drugs fail during development - resulting in losses of around 170 billion euros. The Graz-based company Innophore is now putting an end to this with artificial intelligence. The new application can predict the effects and side effects of drugs and is already finding customers in the USA.

The analysis of human genes was the starting point: researchers from the Graz-based tech-bio company Innophore examined their three-dimensional structure together with the team from Nvidia in San Francisco. The result was a groundbreaking kit: "This is the most comprehensive structural data set on the human organism currently available worldwide," explains Christian Gruber, scientist and CEO of Innophore.

Christian Gruber (third person from right) and his team at Innophore.

(Bild: Innophore GmbH)

But what do these kits tell us? They facilitate the prediction of protein functions and can train AI applications on how drugs work. "The collaboration has led to more than half a million potential binding sites for drugs," says David Ruau from Nvidia. The resulting artificial intelligence can now help to identify the side effects of drugs at an early stage of development.

Artificial intelligence should save costs
This is a major advantage, as almost 90 percent of drugs developed fail - often due to side effects. This results in annual losses of 140 billion euros in development and 30 billion euros in clinical follow-up costs.

The new AI tool is already being integrated into Innophore's research and ensures more effective drugs. An international pharmaceutical company in the USA is also already using the tool. But the field of application could be much wider: Because artificial intelligence is currently being developed to also create climate-friendly processes in industry. These results - which originated in Graz - have now been published in the renowned journal "Nature - Scientific Data".