Artificial intelligence (AI)-driven researchers have created a novel machine-learning system that is capable of foreseeing the appearance of novel viral variations with important features and closely monitoring the evolution of epidemic viruses.
The researchers used data on recorded SARS-CoV-2 variations and Covid-19 death rates to establish the system’s effectiveness in a study that was published in Cell Patterns.
Even before the World Health Organisation (WHO) formally recognised them as such, the AI-based technique was successful in predicting the emergence of new SARS-CoV-2 Variants of Concern (VOCs). This innovation makes it possible to track viral pandemics in real time in the future.
The possible effects of this machine-learning strategy on public and private health organisations were discussed by William Balch, professor in the Department of Molecular Medicine at Scripps Research Translational Institute in the United States. With the use of this technology, it will be possible to prevent viral outbreaks by understanding the hitherto poorly understood rules of pandemic virus evolution.
The researchers focused on the Covid-19 pandemic, employing Gaussian process-based spatial covariance to relate three key datasets: genetic sequences of SARS-CoV-2 variants found worldwide, variant frequencies, and global mortality rates for Covid-19.
To add more, they used this technique to track genetic alterations in SARS-CoV-2 variants worldwide. These modifications, which showed higher dissemination rates and lower mortality rates, showed how the virus had adapted to numerous circumstances, including lockdowns, mask use, immunisations, and innate immunity.
The approach found important gene variants that increased in prevalence as mortality rates changed, all of which occurred prior by several weeks to the WHO’s formal identification of VOCs carrying these variants.
One of the key conclusions from the study stressed the significance of taking into account both prominent variants and the enormous number of undesignated variants, commonly known as the “variant dark matter.” This thorough method enables a more thorough comprehension of viral evolution.
Real-time monitoring and pandemic forecasting are two other potential uses for this surveillance system. Health officials can quickly put in place the necessary countermeasures by foreseeing changes in a pandemic’s trajectory, such as sharp spikes in infection rates.
Additionally, the researchers envision leveraging this AI-driven methodology to gain deeper insights into virus biology, ultimately enhancing the development of effective treatments and vaccines.