AI Increases COVID-19 Vaccine Antibody Response by 128X

Baidu Research, a Chinese tech company, has developed an AI algorithm that could design stable COVID-19 mRNA vaccine sequences in a matter of minutes.

The algorithm, called LinearDesign, has been able to achieve a 128-fold increase in antibody response for COVID-19 vaccine sequences, according to a new study.

COVID-19 Omicron Booster Shots Administered Along With Other Vaccines At Chicago Senior Center — CHICAGO, ILLINOIS - SEPTEMBER 09: A pharmacist prepares to administer COVID-19 vaccine booster shots during an event hosted by the Chicago Department of Public Health at the Southwest Senior Center on September 09, 2022 in Chicago, Illinois. The recently authorized booster vaccine protects against the original SARS-CoV-2 virus and the more recent omicron variants, BA.4 and BA.5. Scott Olson/Getty Images

Lattice Parsing

The natural instability of mRNA makes it challenging to store and transport vaccines, which is especially difficult in developing countries with limited resources.

Scientists wanted to make better vaccines by improving the stability of the mRNA that tells cells how to make proteins. However, there are so many possible variations of mRNA that can make the same protein that it's hard to find the best one.

To solve this problem, they used a method called lattice parsing, which creates a graph to represent all the possible mRNA sequences. Using this method, they found the best mRNA sequence to make the Spike protein.

The new mRNA vaccine sequences designed using this method were better than the existing ones. For COVID-19 vaccines, the new sequences were up to 5 times more stable, produced 3 times more protein, and triggered 128 times more antibodies. For VZV vaccines, the new sequences were up to 6 times more stable, produced 5.3 times more protein, and triggered 8 times more antibodies.

"The vaccines designed through our method may offer better protection with the same dosage and potentially provide equal protection with a smaller dose, leading to fewer side effects," said Dr. He Zhang, Staff Software Engineer at Baidu Research.

"This will greatly reduce the vaccine research and development costs for biopharmaceutical companies while improving the outcomes."

Baidu's AI Ventures

PaddleHelix is a bio-computing platform based on PaddlePaddle that incorporates the ERNIE-Bio-Computing Big Models developed by Baidu. It applies AI to different areas, including small molecules, proteins/peptides, and RNA, providing a new research realm for AI in life sciences.

The ERNIE Big Model from Baidu has created an extensive big model technology system covering NLP, vision, cross-modal, and bio-computing.

The ERNIE Bot is part of the ERNIE Big Model family and is a large knowledge-enhanced language model (LLM) that has the capability to comprehend and produce human language. Sanofi and Baidu formed a partnership in 2021 to integrate the LinearDesign algorithm into Sanofi's mRNA vaccine and drug development pipeline.

"This research can apply mRNA medicine encoding to a wider range of therapeutic proteins, such as monoclonal antibodies and anti-cancer drugs, promising broad applications and far-reaching impact," Dr. Zhang added.

This study is notable for being the first time in which a Chinese technology company is listed as the primary affiliation in a paper published in the journal Nature.