Science Daily September 28, 2021
Researchers in the UK developed machine learning models that identify candidate zoonoses solely using signatures of host range encoded in viral genomes. Within a dataset of 861 viral species with known zoonotic status, their approach outperformed models based on the phylogenetic relatedness of viruses to known human-infecting viruses distinguishing high-risk viruses within families that contain a minority of human-infecting species. The model predictions suggested the existence of generalizable features of viral genomes that are independent of virus taxonomic relationships and that may preadapt viruses to infect humans. Their model reduced a second set of 645 animal-associated viruses that were excluded from training to 272 high and 41 very high-risk candidate zoonoses. A second application showed that their models could have identified SARS-CoV-2 as a relatively high-risk coronavirus strain and that this prediction required no prior knowledge of zoonotic severe SARS-related coronaviruses. Genome-based zoonotic risk assessment provides a rapid, low-cost approach to enable evidence-driven virus surveillance and increases the feasibility of downstream biological and ecological characterization of viruses…read more. Open Access TECHNICAL ARTICLEÂ
Machine learning prediction of human infectivity from viral genomes. Credit: PLOS Biology, September 28, 2021Â