Predicting Plagues: A Call to Create a Predictive Model for Zoonotic Development

Abstract

Zoonotic viruses are becoming more common and oftentimes result in epidemics, like the Ebola virus, or pandemics, like COVID-19. Recent increases in human-animal contact are a systemic problem unlikely to change. As contact increases, it is important to determine what diseases are most likely to spill over. This early warning sign will allow for early vaccine development and prevention measures to be put in place. Bats form the most likely reservoir for emerging zoonoses as they have increased contact with humans and intermediate hosts, act as natural reservoirs for coronaviruses, and have close evolutionary relationships with humans when compared to other common viral vectors. Coronaviruses themselves have shown an increased ability to spillover causing several widespread outbreaks in less than twenty years. Coronaviruses also have many viral characteristics that increase their zoonotic potential such as high mutation rates. For a virus to change hosts, its receptor binding domain (RBD) must be able to recognize and bind the new host's cell receptors. These changes could be tracked to determine the likelihood of the virus to bind to human receptors. Here, I propose that a model can be developed to predict which viruses are more likely to spill over and when. There is increasing amounts of research that discuss the various evolutionary constraints and processes that viruses go through as well as the characteristics of host and virus that increase spillover potential. With use of a machine learning model, the rate and types of mutations can be estimated for the receptor binding domain or various diseases. These models will need to be trained to account for host and viral factors that influence mutations and the evolutionary constraints exerted on zoonotic viruses such as the founder effect. This will allow for the prediction of zoonotic outbreaks.