Jun 28 2021
A new study performed at the University of Liverpool could help researchers alleviate the future spread of livestock and zoonotic diseases caused by viruses.
Abstract representation of networks of (1) observed and (2) predicted associations between wild and semi-domesticated mammalian hosts and known virus species. Image Credit: Dr Maya Wardeh.
A so-called form of artificial intelligence (AI) known as machine-learning has been utilized by scientists to forecast over 20,000 unidentified links between vulnerable mammalian species and familiar viruses. The study results reported in the journal Nature Communications could help target disease surveillance programs.
Mammals have been affected by thousands of viruses, with new evaluations indicating that below 1% of mammalian viral diversity has been found so far. A few of these viruses, like human and feline immunodeficiency viruses, possess a highly narrow host range, whereas others like West Nile and rabies viruses have host ranges that are broader.
Host range is an important predictor of whether a virus is zoonotic and therefore poses a risk to humans. Most recently, SARS-CoV-2 has been found to have a relatively broad host range which may have facilitated its spill-over to humans. However, our knowledge of the host range of most viruses remains limited.
Dr Maya Wardeh, Study Lead Researcher, Institute of Infection, Veterinary and Ecological Sciences, University of Liverpool
For this knowledge gap to be bridged, a novel machine learning framework has been developed by scientists to forecast unfamiliar links between familiar viruses and vulnerable mammalian species by combining three clear perspectives — that of each mammal, each virus and the network that links them.
The study's findings indicate that there are over five times as many connections between well-known zoonotic viruses and wild and semi-domesticated mammals compared to what was originally thought.
In particular, rodents and bats, which have been linked with current outbreaks of emerging viruses like hantaviruses and coronaviruses, were associated with a higher risk of zoonotic viruses.
Moreover, the model forecasts a five-fold increase in connections between wild and semi-domesticated viruses and mammals of economically significant domestic species like livestock and pets.
As viruses continue to move across the globe, our model provides a powerful way to assess potential hosts they have yet to encounter. Having this foresight could help to identify and mitigate zoonotic and animal-disease risks, such as spill-over from animal reservoirs into human populations.
Dr Maya Wardeh, Study Lead Researcher, Institute of Infection, Veterinary and Ecological Sciences, University of Liverpool
At present, Dr. Wardeh is extending the method to forecast the potential of insects and ticks to transmit viruses to mammals and birds, which will allow prioritization of laboratory-based vector-competence studies throughout the world to help avoid outbreaks of vector-borne diseases in the years to come.
Journal Reference:
Wardeh, M., et al. (2021) Divide-and-conquer: machine-learning integrates mammalian and viral traits with network features to predict virus-mammal associations. Nature Communications. doi.org/10.1038/s41467-021-24085-w.