Nearly 2 million animal viruses can infect humans
According to researchers, artificial intelligence (AI) could help scientists identify the next animal virus that is capable of infecting humans.
In a study published Tuesday in the journal PLOS Biology, the Glasgow-based team said it has created a genomic model that can “retrospectively or prospectively predict the likelihood that viruses will be able to infect humans.”
Group Developed Machine Learning Models To exclude candidate zoonotic viruses using the signature of the host range encoded in the viral genome.
GOOGLE KNOWS WHAT YOU TYPE, SEE AND SAY — UNLESS YOU TAKE THIS STEP
With a dataset of 861 viral species with known zoonotic status, the researchers collected a single representative genome sequence from hundreds of RNA and DNA virus species spanning 36 viral families.
They classified each virus as being able to infect humans or not, created by merging three previously published datasets that reported data at the virus species level and host range within virus species. did not consider the possibility of variation.
The researchers trained the models to classify the virus accordingly.
Binary predictions correctly identified about 72% of viruses that primarily or exclusively infect humans and about 70% of zoonotic viruses as infecting humans, although performance varies between viral families.
On further converting the predicted probabilities of zoonotic potential into four categories, 92% of human-infecting viruses were predicted to have moderate, high or very high zoonotic potential and a total of 18 viruses are currently known to infect humans by their criteria. was not estimated for. There is much zoonotic potential – at least three of which had serological evidence of human infection, suggesting that they may be valid zoonoses.
“In the full dataset, 77.2% of the predicted viruses were known to have very high zoonotic potential to infect humans,” the researchers wrote.
Next, the scientists tested several learning-based models to find the best-performing model, which was used to rank 758 virus species — and 38 viral families — that were not present in the training data. .
In the second set of viruses associated with 645 animals excluding training data, the model predicted an increased risk of zoonotic transmission of genetically identical nonhuman primate-associated viruses.
Unless you change this setting, your apps are visible everywhere you go
“Taken together, our results are consistent with the expectation that the relatively close phylogenetic proximity of nonhuman primates may facilitate virus sharing with humans and suggest that this is partly due to the viral genome in both humans and nonhuman primates. may reflect common selective pressures on structure. However, wide differences between other animal groups appear to have less impact on zoonotic potential than virus characteristics,” the authors said.
Overall, 70.8% of viruses from humans with high or very high zoonotic potential were correctly identified.
A second case study predicted the zoonotic potential of all currently recognized coronavirus species and the human and animal genomes of all severe acute respiratory syndrome-related coronaviruses.
“Our findings suggest that the zoonotic potential of viruses can be inferred to a surprisingly large extent from their genome sequences,” the researchers explained. “By highlighting the viruses with the greatest potential to become zoonotic, the genome-based ranking allows further ecological and virological characterization to be targeted more effectively.”
By identifying and further testing high-risk viruses, he said the predictions could aid in the growing imbalance between the rapid pace of virus discovery and the research needed to comprehensively assess the risk.
About 2 million animal viruses can infect humans.
“Importantly, given the clinical limitations and the possibility that not all viruses capable of human infection have had opportunities to emerge and be detected, viruses that are not reported to infect humans are unrealistic, may represent unspecified, or truly non-zoonotic species. Identifying potential or undescribed zoonoses within our data was a primary goal of our analysis,” the group said.
“A genomic sequence is usually the first, and often only, information we have on a newly discovered virus, and the more information we can extract from it, the sooner we can identify the virus’s origin and zoonotic risk. ” Simon Babayan, author of the Biodiversity Institute at the University of Glasgow said in a journal news release.
“As more viruses are characterized, our machine learning models will be more effective in identifying rare viruses that should be closely monitored and prioritized for preemptive vaccine development,” he said.