Scientists from The Universities of Manchester and Oxford have created an artificial intelligence (AI) tool designed to better spot and track the emergence of new COVID-19 variants, a breakthrough that might also be applied to other infectious diseases in the future.
This new method could make it easier and faster to identify virus strains that could pose a significant risk, enhancing our ability to respond to pandemics.
The core of this innovation is a combination of advanced data analysis techniques and a newly developed clustering algorithm known as CLASSIX, which was created by mathematicians at The University of Manchester.
This approach enables the rapid grouping of viral genomes, potentially harmful ones, from a vast array of data.
The research, shared in the journal PNAS, shows promise in augmenting traditional viral tracking methods, such as phylogenetic analysis, which typically demands a lot of manual effort and time.
Roberto Cahuantzi, a researcher at The University of Manchester and the lead author of the study, highlighted the urgency of detecting worrisome new COVID-19 variants like alpha, delta, and omicron.
These variants have been responsible for multiple waves of the pandemic, featuring increased transmission rates, immune response evasion, and more severe illness.
The development of this AI framework could significantly speed up the detection process, aiding in more proactive responses, including specialized vaccine development and potentially stopping variants in their tracks before they spread widely.
The challenge lies in the nature of COVID-19 and similar RNA viruses, which mutate quickly due to their high mutation rates and short generational spans.
This rapid evolution makes it difficult to pinpoint future problematic strains among the millions of sequences available in databases like GISAID, which hosts genomic data for influenza viruses and now includes nearly 16 million COVID-19 sequences.
Traditional methods of mapping the evolution of all these COVID-19 genomes require significant computational and human resources.
The new method proposed by the team could automate these tasks, processing millions of high-quality sequences in a matter of days on a standard laptop—a task that would be unfeasible with existing methods.
This efficiency opens up the possibility for more researchers to identify concerning pathogen strains without the need for extensive resources.
Thomas House, a professor involved in the study, emphasized the importance of improving our methods to keep up with the rapidly growing amount of genetic data produced during the pandemic.
The aim is not to replace human expertise but to complement it, allowing experts to focus on other critical tasks.
The AI method simplifies genetic sequences into smaller segments, or “words,” and uses machine learning to group similar sequences together.
This process, powered by the CLASSIX algorithm, is less demanding computationally than traditional methods and provides clear explanations for its findings.
This proof-of-concept demonstrates how machine learning can serve as an early warning system for the emergence of major new virus variants, offering a scalable and low-cost alternative to phylogenetic analysis.
While phylogenetics remains the definitive method for understanding viral ancestry, the new AI approach can handle vastly more sequences, making it a valuable tool in our ongoing battle against pandemics.
If you care about dementia, please read studies about Scientists find a drug related to Viagra may help treat vascular dementia and findings of link between body inflammation and dementia risk.
For more information about brain health, please see recent studies about diet that may help prevent Alzheimer’s, and results showing some dementia cases could be prevented by changing these 12 things.
The research findings can be found in PNAS.
Copyright © 2024 Knowridge Science Report. All rights reserved.
