Are you sure?

Biotech Startup Aims to Redefine Cancer Drug Scene

Fighting cancer might be about to enter a new era thanks to Turbine, a Hungarian startup whose AI-guided solution could radically cut experimenting time for drugs and help design effective combination therapies.

Szabolcs Nagy, CEO of Turbine.

There is probably just one thing harder than treating cancer; namely, getting to the point where the fight can start in the first place with the help of accredited drugs. Hungarian startup Turbine aims to shake things up on the experimental scene; if it succeeds, the cancer drug industry is up for a major breakthrough.

The fact that Turbine pocketed three awards at the 2017 Central European Startup Awards’ (CESA) regional final (Startup of the Year, Best Biotech Startup and Best AI Startup), and was also Health Startup of the Year at Pioneers, should be a more than promising sign. 

As CEO Szabolcs Nagy explains to the Budapest Business Journal, it takes USD 2.7 billion and one decade for any cancer drug to be marketed. The reason for such astronomic costs and lengthy authorization procedure is the nearly 99% failure rate of research. The biggest problem is that the disease can assume thousands of different forms at any time, so molecularly diverse tumors require different treatments.

The Turbine team started to model how cells work in 2010 and have now built the largest known model of simulated cancer cells. “Last year, we got the chance to test this sci-fi tech in action at Bayer, one of the largest pharma companies researching cancer drugs. This year, they asked us to help increase the effectiveness of several drugs,” says the CEO.

Turbine’s innovation is rooted in the fact that it replaces laboratory experiments with AI-based simulation. This way, millions of trials can be run on servers in the time it would take to run just one biological experiment in the real world. The results show the most promising ways to use a drug, which dramatically speeds things up.

Complex and Constantly Changing

“In the case of cancer drug research, you can’t just grab a bunch of data, crunch it with an algorithm and predict what will happen accurately,” Nagy says, referencing to the practice commonly applied by developers working on autonomous driving systems, for example. “We are talking about a complex and constantly changing biological system. Standard artificial intelligence techniques barely produce better predictions than random guessing.”

Turbine’s cell model has been set up by relying on thousands of publications, and its AI tool predicts how drugs impact cancer by running millions of simulations based on that.

“You need to play around with the molecular setup of cells and what drug you administer in what dose to figure out the best way to tackle cancer,” says Nagy, who is anything but a business rookie. Prior to joining Turbine, he was one of the first people in the team that launched Tresorit, the world’s leading cloud security startup, and helped develop Webicina, a platform that helps patients and physicians find reliable and correct advice on medical conditions and therapeutic opportunities.

The Tresorit experience taught him how to put together a great team and make it function, a skill he says he can surely make use of at Turbine, since the staff is a diverse crew of machine learning experts, biologists, bio-informaticians, medical doctors and programmers. “Over the past few years of building Turbine, and by talking with many people in the pharma industry we have gained insight into what there is demand for, and have aimed to develop our solution towards responding to real needs in research,” concludes Nagy.