Springe direkt zu Inhalt

New Substances Found Using Computers

The open-source platform “VirtualFlow” examines billions of molecules in record time

News from Apr 24, 2020

An international team of researchers, including researchers from Freie Universität Berlin and Technische Universität Berlin, has succeeded in developing a platform that can analyze billions of potential active ingredients in an extremely short period of time. The work was published in the renowned scientific journal Nature. Motivated by the outbreak of the coronavirus pandemic, the researchers have now started to use the “VirtualFlow” platform in order to search for potential active substances that could block the coronavirus proteins.

The development of new active substances in medicine can take many years and may cost billions of euros. A tried-and-tested principle of drug research is to inhibit disease-causing proteins by introducing small molecules into the body. These molecules bind themselves firmly to the disease-causing protein at a certain position, thereby changing or inhibiting it. The issue, however, is that there may be a large number of small molecules that could form the basis of a new drug, and each of them would have to be tested. Computer simulations, which replace elaborate laboratory experiments, have the potential to solve these issues.

“In a computer simulation, we can test the ability of a small molecule to bind to a disease-causing protein – this is also called ‘docking.’ If this is done with multiple small molecules, we call this virtual screening. This type of approach already exists, but so far only a relatively low number of small molecules could be routinely tested in such virtual screenings,” explains Christoph Gorgulla, a postdoctoral research fellow at Harvard University, who received a scholarship from the Einstein Center for Mathematics Berlin (ECMath) at Freie Universität Berlin. The major findings of the publication are based on his doctoral thesis.

The team of researchers developed the “Virtual Flow” platform which enables not millions but billions of small molecules to be concurrently tested for their ability to bind to a specific protein in a short period of time by computer simulation. “Instead of an ordinary computer, ‘VirtualFlow’ is powered by supercomputers that contain thousands of processors. It also works on cloud servers,” says Dr. Konstantin Fackeldey, a private lecturer at the Institute of Mathematics at Technische Universität Berlin. Dr. Fackeldey is a member of the project team that cooperates with the Zuse-Institut Berlin within the framework of the Berlin Math+ Cluster of Excellence.

Google has now provided the researchers with additional resources, enabling them to use its cloud computing platform in order to search for possible active substances that can counteract the effects of coronavirus. “VirtualFlow” is an open-source platform available to all researchers free of charge.