When the novel coronavirus SARS-CoV-2 that's causing COVID-19 started dominating news headlines early last year, there was a lot of conjecture about the nature and propensity of the virus to mutate. We're nearly 15 months into the pandemic and several mutations have been officially reported. Upon hearing the word mutation, laymen do tend to panic but scientists have long assuaged people that mutations are not alarming, and certainly not unusual with novel viruses. In fact, every virus mutates - its part of the lifecycle. Most of these changes aren't cause for concern; some of these changes can actually weaken the virus. 

For over a year now, we've all been reading about the virus, mutations, variants and strains. Let's first decode these terms. 

SARS-CoV-2 is a novel coronavirus that causes COVID-19. It's called novel because it has not been identified by humans before and the global scientific community is still learning about this virus and its impact on man as it spreads across the world. The novel coronavirus is an RNA virus i.e. a collection of genetic material packed within a protein shell. Once it enters a host (through the nose or mouth of a human), it starts making copies of itself and multiples within the system. Sometimes, errors occur during this copying process - this is called a mutation. Mutations are common with every virus. Since viruses are always changing, new variants are emerging and being reported across the world. Some of the commonly seen variants include the Brazil variant, the UK variant, the South Africa variant and now, the India variant. Sometimes, the changes a virus undergoes could make it more infectious, as currently seen in countries like India and Brazil. Other changes could also weaken the virus' potency. It just depends on what that change really is. The Indian variant (B.1.167) is called a "double mutant" due to the presence of two such changes in the virus, and researchers are alluding that this variant is even more transmissible than the UK variant. 

So, this virus is changing the game every few weeks or months. Even though vaccines like Pfizer-Moderna and India's Covaxin are supposed to be able to counter these new strains, studies are ongoing to ascertain their ability to provide adequate protection. At such time, the scientific community is relying on technologies like AI to possibly predict and detect mutations in advance. This can help develop more effective vaccines. 

A research team from the University of Southern California Viterbi School of Engineering developed a new model that can counter emergent variants of the novel coronavirus. The AI model can speed up the process of vaccine development by analysing its efficacy. The study stated that the method is easily adaptable to analyse potential mutations of the virus, ensuring the best possible vaccines are quickly identified. The researchers of the study believe that this will give an edge to the science community against evolving contagion. The ML model can accomplish vaccine design cycles that once took months or years in a matter of seconds and minutes, the study stated. The AI-assisted method predicted 26 potential vaccines that would be effective against the coronavirus. From those, the scientists identified the best 11 from which to construct a multi-epitope vaccine, which can attack the spike proteins that the coronavirus uses to bind and penetrate a host cell.

Another essential tool in detecting mutations is NLP. Several properties of biological systems can be interpreted in terms of words and sentences. “We’re learning the language of evolution,” says Bonnie Berger, a computational biologist at the Massachusetts Institute of Technology. Protein sequences and genetic codes can be modelled using NLP techniques, and further, to detect mutations. This study decodes the idea of a viral immune escape through mutation and how NLP could be a key tool in detecting this genetic sorcery. 

Berger's team used grammar and semantics to identify a virus - a successful virus is grammatically correct while an unsuccessful one is not. Mutations can be detected through semantics - viruses with different mutations can have different meanings, and a virus with a different meaning may need different antibodies to read it. To model these properties, the researchers used an LSTM, a type of neural network that predates the transformer-based ones used by large language models like GPT-3. These older networks can be trained on far less data than transformers and still perform well for many applications, according to this report in MIT. NLP models work by encoding words in a mathematical space in such a way that words with similar meanings are closer together than words with different meanings. This is known as embedding. For viruses, the embedding of the genetic sequences grouped viruses according to how similar their mutations were. 

Using these tools accelerates the process of detecting virus behaviour - and time is of essence in this ongoing battle against COVID19. 

Want to publish your content?

Publish an article and share your insights to the world.

ALSO EXPLORE

DISCLAIMER

The information provided on this page has been procured through secondary sources. In case you would like to suggest any update, please write to us at support.ai@mail.nasscom.in