AI, virtual reality can help fast track COVID-19 vaccine, experts say

From helping in optimising the yield of therapeutics to training staff for setting up large-scale manufacturing sites, cutting-edge technologies such as artificial intelligence (AI) and virtual reality (VR) can be used to fast track COVID-19 vaccine development worldwide, experts say.

The search for a COVID-19 vaccine has expanded worldwide, with thousands of researchers collaborating at hundreds of laboratories to fight the virus that has infected 56 million people and claimed over 1.34 million lives so far.

Recently, a panel of experts noted at the Berlin Science Week, a ten-day science festival, that AI and other technologies like machine learning (ML) can make sense of the mountains of data from several experiments by discovering patterns that a human brain might fail to spot.

As vaccine candidates advance to the final phases of testing in humans, experts said AI would be vital for analysing clinical and immunological data rapidly.

Rene Faber, from the pharmaceutical company Sartorius headquartered in Germany, said there is a need to utilise these "handy innovations."

Citing the example of single-use technologies, Faber said these can help build manufacturing capacities for the vaccines quickly.

In single-use technologies, he said, everything in the manufacturing process is made out of plastic, and is used just once to help users avoid any cleaning.

"It helps customers to build manufacturing facilities much faster, saving a year or even more than a year to build the manufacturing process, Faber said during a session on current trends in vaccine research organised by the Falling Walls Foundation.

"And also it's something which is very flexible, meaning if you build such a facility, let us say in the US, you can copy-paste very quickly that manufacturing process either in Europe or in Asia as well," he noted.

Another technology of importance in COVID-19 vaccine research, Faber said, is automation.

When customers develop a manufacturing process, he said they have to run a large number of experiments to narrow in on the right process parameters.

According to Faber, manufacturers also need to understand how the production process can impact the quality of the vaccine, and find ways to optimise its yield.

"This is very important work which needs to be done and takes a lot of time, and there are very innovative tools which allow for very high throughput, and automation is a way to run such experiments," he said.

Digitisation, and tools such as AI are used for real-time data analysis to predict or change the process before something goes wrong, he added.

"And you may avoid loss of a vaccine badge which of course nobody wants to have in the situation today. So there are a number of innovations we have in our hands today that we can use to really speed up the development and also build up the manufacturing," Faber said.

Uwe Gottschalk, Chief Technology Officer at Lonza Pharma/Biotech, Switzerland said in the therapeutic space, especially vaccine development, new innovations and technology are coming to help.

" the pharma industry as a whole is quite slow and that is because we cannot cut corners and compromise the efficacy and safety of our products, and this has also been for too long used as an excuse not to innovate, and from gene to R&D the mantra was it takes one or two years to make clinical material from the DNA becoming available for a new drug, Gottschalk said.

With this being challenged in the vaccine industry, he said companies are now applying all the know-how from ML to AI for speeding up vaccine development.

Citing another example of novel technology coming to aid amidst the COVID-19 pandemic, Gottschalk said Lonza is currently setting up large-scale manufacturing sites for vaccine production even before any equipment is on-site.

"We can train our specialists with the help of virtual reality. So the specialists are being trained as we speak. Although they are sitting in a training centre or maybe at home," he said, adding this approach can help save months.

Faber said the production of most biological products is based on the behaviour of individual cells, and AI is used to model how these building blocks of life act as a manufacturing facility for proteins and other molecules, including vaccines.

"Tools like AI, mechanistic modelling, statistical tools, statistical data analysis, all in combination we believe will bring a lot of insights in the challenges we still have today," he added.

Highlighting the use of AI in vaccine development, Professor Carlos A. Guzman, head of the department Vaccinology and Applied Microbiology at the Helmholtz Centre for Infection Research, Germany said every individual human even within the same age group does not respond to vaccines in the same manner, adding that it is important to understand why there are good responders and bad responders to the therapeutic.

Guzman believes that if the mechanism behind this is understood, novel vaccines can be developed which can even work for groups of the population that generally have poor response to immunisation.

"This kind of technology of course can be implemented and is something we are getting from other fields," he added.