Computer engineers at the University of Southern California have produced a deep learning framework for very quickly fine-tuning vaccines to fight emerging variant strains of COVID-19.
The approach uses in silico modeling, which simulates biological processes on computers, to whittle down large numbers of vaccine candidates to the few most likely to succeed.
Corresponding author Paul Bogdan, PhD, and colleagues at USC’s Viterbi School of Engineering describe the work in a study published Feb. 5 by Nature Research’s Scientific Reports.
In testing and validation, the AI framework—dubbed DeepVacPred by its creators—eliminated 95% of the starting field and plucked out 26 promising vaccine variations in less than a second.
This allowed the team to skip the most labor-intensive part of in silico vaccine design, the authors note.
Next they narrowed the 26 to the 11 best for building a multi-epitope vaccine, the kind needed to take out the spike proteins now so familiar in artistic renderings.
“With DeepVacPred, a researcher can construct a multi-epitope vaccine for a new virus and validate its quality within an hour,” Bogdan and co-authors write.
Applied to the particulars of the SARS-CoV-2 virus, DeepVacPred “can provide vaccine candidates within seconds and move them to clinical trials quickly to achieve preventive medical therapies without compromising safety,” Bogdan tells USC News.
What’s more, he adds, the approach “can be adapted to help us stay ahead of the coronavirus as it mutates around the world.”
The study is available in full for free.
Dave P. has worked in journalism, marketing and public relations for more than 30 years, frequently concentrating on hospitals, healthcare technology and Catholic communications. He has also specialized in fundraising communications, ghostwriting for CEOs of local, national and global charities, nonprofits and foundations.