A model for predicting coronavirus mutations and variants

A model for predicting coronavirus mutations and variants

The computational model makes it possible to predict the appearance of spike protein mutations that could make Sars-Cov-2 variants more contagious or dangerous

(photo: Getty Images) The hunt for new coronavirus variants does not stop Firm: It's important to keep your guard up with virus tracing, case tracking, and genome sequencing. In addition to discovering the mutated forms of the virus early, as soon as they appear, a new tool, still under study, could in the future allow us to be able to predict changes, that is, new mutations, in the Sars-Cov-2 genome. These mutations could give rise to variants that are more contagious and elusive to the immune system, therefore more dangerous. Developing the new computational model, based on complex simulations, is a group from Penn State University, which has published the results in the journal Proceedings of the National Academy of Sciences (PNAS). The idea is to anticipate the process of detecting changes in the virus by preventing its moves.

Artificial intelligence comes to our aid

Scientists started by studying the behavior of the virus and from the information already known on how this changes to build a computational model that is able to anticipate and predict its moves. To obtain the model, they combined and reworked the data using artificial intelligence algorithms, in particular deep learning (field of machine learning). The accuracy of the method, therefore the precision with which, within the database already collected and known, it is able to predict significant alterations of the virus is higher than 80%, the authors write in the text, therefore "rather effective".

The tool also allows to predict the mutations of amino acids (which make up the protein) already observed in some variants, such as alpha, beta and delta. This result could also help predict future changes and possible evolutions into new, more contagious and dangerous forms (remembering that fortunately all variants are currently covered by the double dose vaccination).

Beware of the spike protein

The researchers have turned their attention to the mechanisms and the moment in which the virus enters our organism: it manages to penetrate through a specific part of it, the spike protein, which like a hook that binds to our cells. The coupling takes place via another protein, the Ace2 receptor, present in some human cells. Researchers focused on the spike protein and how it intercepts and connects to Ace2, allowing virus entry and infection.

Specifically, they placed the part of the spike, called receptor binding domain (Rbd), the characteristics of which are central to understanding if and with what ability the entry of the virus occurs and into human or animal cells. Current Sars-Cov-2 variants in circulation all have one or more mutations that have led to changes in this part of the coronavirus protein, the authors explain, which may have conferred greater transmissibility and an increased ability to infect.

"The strength of the link between the Rbd domain and the Ace2 receptor - underlines Suresh Kuchipudi, professor of veterinary medicine and biomedical sciences at Penn State - has a direct effect on the dynamics of the infection and on the progression of the disease" . The ability to predict with high probability how the virus will change and how this will have an effect are important tools, continues the author, for the evaluations of specialists also of possible new spillovers and of how the pathogen is adapting in humans and animals.


Cinema - 3 hours ago

The apocalypse: 10 films to recover about catastrophes , virus and ... the end of the world


What do we know about myocarditis cases reported after vaccines


Time to talk about pandemic depression

Topics

Coronavirus Artificial intelligence Coronavirus vaccine Sars-Cov-2 variants globalData.fldTopic = "Coronavirus, Artificial intelligence, Coronavirus vaccine, Sars-Cov-2 variants "

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivs 3.0 Unported License.




Powered by Blogger.