ČT 24: Brno researchers use artificial intelligence to investigate thousands of possible anti-coronavirus drugs
Researchers from Brno have proposed several existing and approved drugs that could prevent the coronavirus binding to human cells. The drugs were identified while researching proteins using artificial intelligence, which also has the potential to identify new drug structures.
The researchers performed a virtual screening of 4359 approved drugs to determine their effectiveness on a protein that is key to the spread of SARS-CoV-2. “The simulations used several conformations of the protein’s molecular arrangement to determine which of the known drugs would have the greatest efficacy," said Dr. Gaspar Pinto of Masaryk University’s Loschmidt Laboratories and St. Anne’s University Hospital.
As the procedure generates an enormous amount of data, machine learning methods and artificial intelligence are required for their analysis.
Based on their calculations, the researchers proposed several approved drugs that could block the function of the protein and thus prevent the virus from binding onto the human host cell. “Artificial intelligence can also help identify new drug structures that bind to proteins even more effectively. This is a new area of covid-19 research, in which software solutions are being developed to accelerate the development of new drugs," said Dr. Pinto.
Artificial intelligence tools
The project, which involved researchers from the Loschmidt Laboratories of the Faculty of Science, the Institute of Computer Science, the RECETOX research centre (all Masaryk University) and the International Centre for Clinical Research at St. Anne’s University Hospital in Brno (FNUSA-ICRC), used computer biochemistry, artificial intelligence and machine learning principles.
“Using our own CaverDock software, we focused on one of the essential proteins responsible for spreading the SARS-CoV-2 virus in the human body," said Dr. Jiří Damborský of the Faculty of Science at Masaryk University, who is also head of the FNUSA-ICRC Protein Engineering research team.
“This project represents the first time that the CaverDock program has been used to study such a large number of molecules. The program performed excellently with practically 100% robustness, making it one of the most reliable tools in its category," said the author of the algorithms used, Dr. Jiří Filipovič of Masaryk University’s Institute of Computer Science.
Translated from https://bit.ly/ct24-koronavirus.
