A new computational model recently mined an unprecedented amount of data, over 6.4 million SARS-CoV-2 sequences, to locate patterns among mutations that help a new strain spread all over the world.

As specified in a Scientific American report, despite having only been existing for less than three years, the COVID-19-causing virus SARS-CoV-2 is probably the most investigated and genetically sequenced pathogen in history.

Disease surveillance teams globally have uploaded millions of viral sequences to public databases that enable scientists to track how the virus is spreading.

The model, known as PyR0, examined how different viral lineages occur and transmit from December 2019 to January 2022.

ALSO READ: Carbon Footprint Measurement Through AI: Scientists Develop New Tool to Track, Calculate CO2 Emissions and Reductions

Coronavirus Variant
(Photo : Wikimedia Commons/ NIAID)
Transmission electron micrograph of a SARS-CoV-2 virus particle (UK B.1.1.7 variant), isolated from a patient sample and cultivated in cell culture.


The PyR0 Model

From the said data, it learned how to determine the combinations of mutations and the amount of time needed for variants like Delta and Omicron to turn prevalent.

The model, which the research team described in the Science journal could provide public health programs with advance notice about which lineage is possibly dangerous and enable officials to plan ahead.

The PyR0 model employed data leading up to mid-December 2021 to properly forecast the BA.2 subvariant of the Omicron strain, which was rare in much of the world at that time, would soon spread fast.

By March this year, BA.2 had turned into the dominant variant worldwide. If the model had been run in November 2020, it would have properly predicted as well, that the Alpha strain would soon turn out to be dominant. The World Health Organization did not identify the said strain as a variant of concern until December of the same year.

Discovering Spike and Non-Spike Protein Mutations

 A World News Era report specified that the PyR0 model discovered that merely having several spike protein mutations did not essentially make a strain more evolutionarily suitable.

However, a few specific spike mutations late last year helped the Omicron subvariants BA.1 and BA.2 escape the immune system.

The said tool also discovered that a set of non-spike mutations in the genome of the BA.2 that impact the manner the virus is replicating might contribute to its fast transmission.

PyR0's ability to rapidly examine entire genomes, the study authors said, might help scientists know which areas of the genome of the virus to investigate in order to develop treatments in the future.

Use of AI in Addressing COVID-19 Crisis

This is not the first time artificial intelligence is used when dealing with COVID-19. In 2021, according to a BBC News report, a team of scientists used AI to work out where the next coronavirus could occur.

The research team used a combination of fundamental biology and machine learning. Their computer algorithm forecasted a lot more potential hosts of new virus variants than have formerly been identified.

In this work, the team asked their AI tool to find use biological patterns to predict which mammals might be vulnerable to known coronaviruses, which showed associations between over 400 strains of coronavirus and over 870 potential mammal species.

Related information about artificial intelligence helping to detect the next coronavirus outbreak is shown on Down to Earth's YouTube video below:

 

RELATED ARTICLE: Pet Owners Can Easily Transfer COVID-19 Infection to Their Cats, Dogs [Study]

Check out more news and information on Artificial Intelligence and COVID-19 in Science Times.