Machine learning algorithms can identify areas of coronavirus genetic makeup that could be the focus of potential treatments. Image The method can predict regions of the coronavirus DNA that could be targeted by treatments. Regions of the coronavirus genome that could be targeted to disrupt the virus can be predicted by a new method. The findings could help to develop vaccines and drug treatments against the virus. Led by Dr Mazdak Salavati from the Roslin Institute, the team who developed the method received the third prize in a coronavirus hackathon for their work. Genetic variations Using genomes publicly available, the method compared the coronavirus between host species (bats and humans) as well as German Covid-19 patients and the first patients in Wuhan, China. The method was developed to identify variations in the virus genome, including common mutations, variations with severe consequences for the virus, and variations specific to the virus when it infects bats, pangolins or humans. Understanding the context of these variations being introduced into the virus genome can help to devise strategies against its spread from one species to another and explain some of the differences in clinical manifestations of Covid-19. The locations of these variations in the virus genome, as well as their function, can be predicted by the new method. Coronavirus hackathon The five-day online CoronaHack event, organised by training and consultancy firm Mindstream AI and life sciences software company DataBiology, brought together data scientists, biomedical researchers and health professionals to work on coronavirus research. The team of bioinformaticians, who had not worked together before, was a collaboration between Dr Mazdak Salavati and Dr Barbara Shih from Roslin, Dr David Parry from the University of Edinburgh’s MRC Institute of Genetics and Molecular Medicine, and Nicholas Dimonaco from Aberystwyth University’s Institute of Biological, Environmental & Rural Sciences. They received a £500 prize and computational resources on the DataBiology platform. The CoronaHack event allowed us to create our own team of bioinformaticians and develop a new method to predict important variations in the coronavirus genome. We thank the organisers and our universities for giving us the opportunity to contribute to tackle the current pandemic. Dr Mazdak SalavatiCore Scientist, the Roslin Institute Congratulations to Maz and team on their success. Innovative projects like theirs are a welcome addition to the Roslin Institute’s ongoing work to tackle Covid-19. Professor Bruce WhitelawInterim Director, the Roslin Institute The team is still working together to apply the method to other available coronavirus genomes. Their study is expected to be published in the coming weeks. ** The Roslin Institute receives strategic investment funding from the Biotechnology and Biological Sciences Research Council and it is part of the University of Edinburgh’s Royal (Dick) School of Veterinary Studies. ** Related links Obesity impacts on Covid-19 recovery prospects Covid-19 tracker helps visualise cases in Scotland CoronaHack — AI vs.Covid-19 #CORONAHACK closing ceremony German Covid-19 patient dataset Image credit: geralt on Pixabay