Photo credit: www.sciencedaily.com
New AI Tool Aims to Combat Pandemic Risks Through Viral Prediction
A groundbreaking artificial intelligence tool has been developed to help mitigate the risk of pandemics by identifying animal species that could harbor and transmit viruses capable of infecting humans.
This innovative machine learning model, created by researchers at Washington State University, evaluates host characteristics and the genetic structure of viruses to pinpoint potential animal reservoirs and geographical regions where new outbreaks may be more likely. The focus of this research is on orthopoxviruses, which include the causative agents of diseases such as smallpox and mpox.
The findings of this research were recently published in the journal Communications Biology. The study highlights the potential for this model to assist scientists in anticipating emerging zoonotic diseases and suggests its adaptability for studying other viral threats.
“Nearly three-quarters of emerging viruses that infect humans originate from animals,” explains Stephanie Seifert, an assistant professor at the WSU College of Veterinary Medicine’s Paul G. Allen School for Global Health and a key leader in the project. “Improving our ability to predict which species are most at risk allows us to enact preventative actions against pandemics.”
According to the model’s predictions, Southeast Asia, equatorial Africa, and the Amazon region emerge as significant hotspots for orthopoxvirus outbreaks. These areas are characterized by high densities of potential host species and correspond with low smallpox vaccination rates. It is noteworthy that the smallpox vaccine offers cross-protection against various orthopoxviruses, yet vaccination efforts ceased following the eradication of smallpox in 1980.
The research indicated several animal families likely to serve as hosts for mpox, including rodents, felids (cats), canids (dogs and related species), skunks, mustelids (e.g., weasels and otters), and raccoons. Importantly, the model successfully excluded rats, which laboratory studies have shown to be resistant to mpox infection.
Katie Tseng, a graduate student in veterinary medicine and the study’s primary author, emphasized that the model not only outperformed previous models in predictive accuracy but also holds promise for identifying hosts for other viruses. “While our model specifically targeted orthopoxviruses, it can be adjusted and applied to a broader range of viral threats,” she noted.
Pilar Fernandez, a disease ecologist and assistant professor in the Allen School who collaborated with Seifert on the project, pointed out that earlier models focused largely on the ecological traits of animals, evaluating factors such as habitat and diet, without considering viral genetics. “While past models effectively analyzed host characteristics, we sought to integrate the genetic attributes of the viruses as well,” Fernandez stated. “Our model enhances host prediction accuracy and provides insights into potential virus transmission across species.”
Orthopoxviruses traditionally result in small, localized outbreaks; however, incidents such as the global surge of mpox in 2022 have raised alarms about these viruses potentially establishing new endemic spots and spreading via new animal hosts.
The identification of possible reservoirs is crucial for predicting spillover events, yet traditional field sampling methods are often resource-intensive and logistically challenging. This new model streamlines the process, allowing for more focused wildlife surveillance efforts.
“In regions like Central Africa, one of the most biodiverse areas on the planet, determining where to start looking for the mpox virus reservoir can be daunting,” Seifert remarked. “By utilizing machine learning models to prioritize sampling efforts, we can more effectively identify the origins of these viruses and gauge the risks they present.”
The research team also included Heather Koehler, an assistant professor in the School of Molecular Biosciences with extensive experience studying mpox. Contributions were made by Daniel J. Becker of the University of Oklahoma, Rory Gibb from University College London, and Collin Carlson of Yale University. These scientists are part of the Viral Emergence Research Institute, a collaborative network focusing on host-virus interactions and virus transmission predictions on a global scale, supported by the National Science Foundation. This team boasts expertise across various fields, including data science, computational biology, virology, ecology, and evolutionary biology.
Source
www.sciencedaily.com