New AI-driven software has written code for a COVID-19 vaccine that provokes up to 128-fold greater antibody responses.
Artificial intelligence (AI) company Baidu Research has unveiled a new algorithm for designing previously impossible mRNA vaccine sequences.
In mouse validation tests, AI-driven software yielded code for a COVID-19 vaccine that elicited up to 128-fold greater antibody responses.
The algorithm, called LinearDesign, has also proven successful in extending the shelf-life of vaccines by a factor of 6, even when exposed to body temperature.
Most vaccines require refrigerated storage, a condition that makes distribution operations difficult, especially in hot regions of the world.
Therefore, if we can extend the shelf stability of vaccines and improve their resistance to different temperatures, it could have a fundamental impact on global vaccination efforts.
Dr. He Zhang, a staff software engineer at Baidu Research, told Euronews Next that the AI tools the researchers have developed have applications beyond vaccines and could help design powerful new cancer treatments. .
Just last week, a personalized mRNA vaccine for pancreatic cancer patients was revealed. promising results In a small study conducted by researchers in New York and BioNTech in Germany, tumor recurrence was prevented in half of treated patients.
The study used pancreatic cancer mRNA vaccines tailored to each patient’s tumor, which may help induce an immune response.
Messenger RNA (mRNA) drugs are a new class of drugs and agents that use small pieces of genetic material called mRNA to teach the body’s cells to make proteins that trigger an immune response against certain pathogens, such as viruses. refers to vaccines. This approach differs from conventional vaccines, which use weakened or inactivated parts of specific pathogens to stimulate the immune system.
Oncology is a key area for mRNA vaccine makers, but allergies is another. BioNTech, the company that produced the COVID-19 (mRNA) vaccine in collaboration with Pfizer, also holds patents for an mRNA vaccination platform designed to protect against allergens such as grass pollen and house dust mites. .
Despite the promise of mRNA medicine, it also comes with its own challenges.
“Although mRNA vaccines have saved many lives, there are still some issues with stability and thus efficacy,” Chan said.
The challenge of keeping mRNA intact
Messenger RNA is unstable because it is single-stranded, unlike double-stranded DNA.
The single-stranded portions of our mRNAs are more easily cleaved or broken into segments by the immune system, water molecules, or even before they are injected into the body, says Oregon. State University professor Dr. Liang Huang explained. Author of the LinearDesign paper.
And when the mRNA breaks down into smaller pieces, it can no longer deliver the complete message to the cell.
“The goal is to keep the mRNA intact,” Huang said, which requires the messenger RNA to be as compact as possible.
“In other words, we want to design mRNAs that are folded and have a more similar shape. [double-stranded] DNA,” he told Euronews Next.
And that’s exactly what the LinearDesign algorithm did.
How does the LinearDesign algorithm work?
A breakthrough algorithm that promises to revolutionize the field of mRNA therapy is actually nothing new.
The LinearDesign model is based on a 1961 algorithm called Lattice Analysis, originally invented for natural language processing and speech recognition.
“We basically used the same algorithm unchanged. The only change is the input, the grammar,” he said.
The lattice parsing algorithm was originally used to identify the most likely sentence among several possible alternatives that sound similar.
“This technology already existed in some form 60 years ago,” Huang said. “Linguistics and biology, so far apart, he was just waiting to discover the link between the two fields.”
Using this technology, researchers have found a way to have an algorithm create the “most stable COVID-19 mRNA vaccine” in just 11 minutes.
In other words, they identified the safest structures for mRNA messengers to move and deliver messages to cells.
Going through all the possible combinations that the AI algorithms did to find the most stable vaccine would take humans “the lifetime of the universe, billions of years,” Huang said.
Will AI be the future of vaccines?
Baidu researchers predict that the LinearDesign algorithm will make mRNA technology “more pervasive than it is today,” allowing other variants of the technology to enter more medical fields.
Paris-based French pharmaceutical giant Sanofi has already licensed the LinearDesign technology. The pharmaceutical company picked up the study in late 2021, when the researchers first published a preprint of the study.
This paper was authored in collaboration with Oregon State University, StemiRNA Therapeutics, and the University of Rochester Medical Center and was published in 2016. Science magazine Nature Huang said the paper was first published under the “uncommon option of early paper preview.” In other words, the scientific journal published the research before it actually underwent the editorial process.
“The editor chose us for a fast preview of the article because, as the quote suggests, ‘we thought this was relevant to public health and the pandemic.'”
Scientists at Baidu Research are optimistic that mRNA technology will replace traditional vaccines within the next decade.
“In about five years, probably all the mRNA pharmaceutical industries will be using this algorithm,” Huang told Euronews Next, adding, “Except for stability, mRNA vaccines are much more It’s been great and we’ve solved it,” he added.
