Zoonotic diseases occur due to fungi or parasites, or bacteria that spread from animal to human. Approximately 60% of known and 75% of new or emerging infectious diseases can spread from animals to humans. According to health experts, the zoonotic disease is an infection that originates in an animal caused by a parasite or pathogen that is happy to live inside the animal.
Occasionally the parasite or pathogen will transfer to humans, and 99% of the time, that is where its ends. The infected person might get sick, but it’s a dead-end host, so it doesn’t go any further. Some of them can transmit from one person to another. Something that has the potential to become pandemic secondary transmission is critical.
Research and surveillance can be improved by identifying high-risk viruses earlier. A study by Daniel Streicker, Simon Babayan, and Nardus Mollentze was published in PLOS Biology on September 28th. They belong to the University of Glasgow, UK. Any animal-infecting virus that will infect humans can be predicted by machine learning using viral genomes if relevant biological exposure is given. Machine learning is a type of artificial intelligence.
Before an emergency, identifying zoonotic disease is a significant challenge because only a tiny minority of the estimated 1.67 million animal viruses can infect humans. Using viral genome sequences to develop machine learning models, a dataset of 861 virus species from 36 families was 1st compiled by the researchers. The researchers then built a machine learning model based on patterns in virus genomes assigned a probability of human infection. The best-performing models were applied by the authors to analyze patterns in the predicted zoonotic potential of additional virus genomes sampled from various species.
It was found by the researchers that viral genomes may have generalizable features. These features are independent of virus taxonomic relationships and may preadapt viruses from infecting humans. The researchers, by using viral genomes, were able to develop machine learning that identifies candidate zoonoses. These models have limitations because identifying zoonotic viruses that can infect humans computer models are only a preliminary step.
Before pursuing significant additional research investments, virus flagged by the models will require confirmatory laboratory testing. Whether humans can be infected by viruses, these models can predict. But the ability to infect is just one part of broader zoonotic risk.
The authors said their findings show –
- From the genome sequence, the zoonotic potential of viruses can be inferred to a large extent.
- Genome-based ranking by highlighting viruses with the most significant potential to become zoonotic allows further virological and ecological characterization to be targeted more effectively.
These findings add an essential piece of information to what can be extracted by the researchers from the genetic sequence of viruses using AI techniques. A genomic sequence is typically the 1st and often only information on the newly discovered virus. The more data the researchers can extract, the sooner they might identify the virus origin and the zoonotic risk it may pose. The machine learning models will be more effective as more viruses are characterized in identifying the rare viruses that ought to be closely monitored and prioritized for preemptive vaccine development.