With the coronavirus covering ground almost across the whole world, scientists and vaccine developers have been worried about the mutations the virus has been projecting, especially recently with at least three different variants of the virus creating panic in countries.
A recent mutation that has been a worrying factor for scientists and doctors alike has been the E484K, having showed up in both the South African variant (B1351) and the Brazilian variant (P1) of the virus. but not in the UK virus strain (B117). Nicknamed the 'escape mutation', it has now shown up in the UK variant of the virus too, but just in 11 cases so far, a report in Scroll.in said.
Rising Number of Mutations
To counter the effects of the mutations resulting in the changing strains, scientists have devised a new method that will help in accelerating the vaccine development process. Researchers from the University of Southern California (USC) have published their paper regarding the same in the journal Nature Scientific Reports.
With at least three variants of the coronavirus swirling around, UK, South Africa and Brazil so far, the scientific world would sure agree for the need to speed up the procedure for vaccine design. This is where the newly designed DeepVacPred comes in, a novel AI-based in silico multiepitope vaccine design framework that combines in silico immunoinformatics and DNN technologies.
The study headed by Zikun Yang, Paul Bogdan and Shahin Nazarian explained in the paper how this AI framework, a deep neural network (DNN) can considerably help in reducing the time required between vaccine design and the subsequent clinical trials.
AI Model to the Rescue
The report explained how the researchers took the help of artificial intelligence (AI) that would understand the intricacies of the vaccine and single out the best option to treat it. The AI model has helped to speed up vaccine design cycles that normally took months and years and shortened the durations to minutes, scientists associated with the study said.
"This AI framework, applied to the specifics of this virus, can provide vaccine candidates within seconds and move them to clinical trials quickly to achieve preventive medical therapies without compromising safety," said Paul Bogdan, associate professor of electrical and computer engineering at the USC Viterbi School of Engineering and corresponding author of the study, in a statement.
How the Model Works
The scientists, in order to run samples through the AI model at first took into account public data related to infectious and such immune mediated disease. The date they required for the SARS-CoV-2 genome and spike protein sequence were taken from the National Center for Biotechnical Information. The model DeepVacPred, was soon able to discard at least 95% of intended compounds and gave out only the most relevant and promising compounds for the vaccine. Within seconds, the AI model was able to suggest 26 potential vaccine subunits that would work against SARS-CoV-2. The researchers then further went on to shorten the list down to 11 compounds that can be used to create a multiepitope vaccine to attack the virus's spike proteins.
Fast Spreading Variant, Quick Solution
The best reason to use the AI model of DeepVacPred is, according to the scientists at the University is that it will be able to build a new multiepitope vaccine in less than a minute and determine its quality within an hour, a process that usually takes upto a year, thus allowing the virus to cause more harm.
A recent study conducted by several scientists have found out that the UK variant of the coronavirus, the first one to emerge end of last year has been spreading in the US quite rapidly. The study found the strain to be doubling its existence among positive covid-19 cases in the US. The CDC in the US had predicted that the UK strain could be the dominant strain in the world by March, thus setting off warning signs everywhere.