Researchers at CERIT-SC, MUNI, and FNUSA-ICRC used their own CoverDock software to test several drugs that could block the virus.
The global pandemic of COVID-19 has affected the lives of millions of people around the world. Health research teams have often radically changed their plans, devoting much of their capacity to research the disease in order to find effective help as soon as possible. Researchers from the Loschmidt Laboratories of the Faculty of Science of Masaryk University (MU), the Institute of Computer Science at MU, the RECETOX MU Research Center and the International Center for Clinical Research at the University Hospital at St. Anny in Brno (FNUSA-ICRC).
The project envisages the use of computer biochemistry, artificial intelligence and the principles of machine learning. "Using our own CaverDock software, we focused on the computer study of a protein that is key in the spread of SARS-CoV-2 virus in the human body," described Jiří Damborský from the Faculty of Science MU and head of the FNUSA-ICRC Protein Engineering research team. It is a viral glycoprotein S whose trimer (a molecule of three monomers) forms the protrusions of the SARS-CoV-2 coronavirus envelope and binds to human host cells.
The researchers performed a so-called virtual screening of 4,359 approved drugs to find out their effectiveness on this particular protein. "In this project, we used the CaverDock program we developed primarily to study such a large number of molecules. The program proved excellent, practically 100% robustness, and thus became one of the most reliable tools in its category, "said the author of the algorithms Jiří Filipovič from the Institute of Computer Science MU (CERIT-SC). The CaverDock program was developed thanks to the support of the MU internal grant agency financing interdisciplinary research and is provided to a wide user community by the national infrastructure ELIXIR CZ.
"We performed several simulations of changes in the molecular arrangement of this protein to see which of the known drugs could have the greatest efficacy," said Gaspar Pinto of Loschmidt Laboratories MU and FNUSA-ICRC. Because a similar process generates an enormous amount of data, machine learning methods and artificial intelligence are used to analyze them. "We also submitted a grant application to Microsoft's Azure cloud program," Pinto added.
Based on these calculations, several approved drugs have been proposed that can block the function of this protein and thus prevent the virus from binding to the human host cell. "Artificial intelligence can also offer new drug structures that bind to protein even more effectively," Pinto said. According to him, this is a new area of COVID-19 disease research, where software solutions are being developed to accelerate the development of new drugs.