This AI has almost exactly detailed the structure and operation of Omicron

An artificial intelligence has helped several researchers to predict the structure at the atomic level of the new variant of coronavirus and, although it is not perfect, it has served to go faster.

Omicron is the strain of COVID that has spread around the world this Christmas. It has always been presented as a much more contagious virus, but milder in its symptoms. Its speed to spread made it a research priority late last year.

Precisely for this reason, the scientific community got down to work. From the University of British Columbia the professor Sriram Subramaniam was among the first to work with genome sequences published online, but received final results obtained by researcher Colby Ford, from the University of Charlotte.

How was it possible for someone to have such quick end results? Subramaniam was one of the first to work on it and from the other side of the country they were getting results. Everything was due to the use of an artificial intelligence that had been invented recently.

While Subramaniam used electron microscopes to reveal the 3D structure of proteins and better understand how they work, Ford had used an AI created from DeepMind AlphaFold between the universities of Alphabet (in the United Kingdom) and that of Washington, to predict how protein peaks would behave.

When Subramaniam had its final results, they looked a lot like what Ford had achieved with prediction software.

Although there were differences, since the program cannot predict how the protein behaves when it comes into contact with certain amino acids, got pretty close to real conclusions.

Ford had only taken an hour to reach conclusions very similar to those of Subramaniam, which had taken two weeks.

“These tools allow you to make an educated guess very quickly, which is important in a situation like Covid.”Says Ford. The AI ​​was not made specifically for the coronavirus and still anticipated the atomic structure of the Omicron variant.

Technology at the service of health

More people submitted their findings based on predictions from artificial intelligence. Although they are not exact, they can be very useful to make the investigation go faster.

By knowing the structure of a protein, you can determine how it behaves. These AIs do not get the behavior right, but they do get a mapping of the structure very close to reality.

It is a complicated task and doing it in such a short time would serve to avoid strenuous laboratory work. Subramaniam hopes that more work will be done with this type of AI in the future. so that the response is faster and more efficient.

It seems that machines could serve much more than enslaving humanity, they could be the ones that save us.