Powerful AI boosts COVID-19 vaccine antibody response by 12,800%

Research has designed a transformative AI algorithm, LinearDesign, which can swiftly produce highly stable COVID-19 mRNA vaccine sequences

[June 10, 2023: Staff Writer, The Brighter Side of News]

When COVID-19 occurred, we of course wanted to study the therapy’s potential and discovered it was effective against every type of coronavirus, in vitro and in vivo. (CREDIT: Creative Commons)

In a groundbreaking development, researchers at Baidu Research have designed a transformative AI algorithm, LinearDesign, which can swiftly produce highly stable COVID-19 mRNA vaccine sequences that were hitherto unattainable.

LinearDesign presents a significant advancement in both stability and efficacy for vaccine sequences, achieving an impressive 128-fold increase in the antibody response of the COVID-19 vaccine.

Dr. He Zhang, Staff Software Engineer at Baidu Research, spoke of the wide-ranging implications of this advancement. “This research can apply mRNA medicine encoding to a broader range of therapeutic proteins, such as monoclonal antibodies and anti-cancer drugs, promising broad applications and far-reaching impact,” he remarked.

The results of this innovative work have been brought to light through a collaborative effort involving Oregon State University, StemiRNA Therapeutics, and the University of Rochester Medical Center.


Related Stories:


The study, titled “Algorithm for Optimized mRNA Design Improves Stability and Immunogenicity,” was featured today in the prestigious scientific journal, Nature, via Accelerated Article Preview (AAP). This instance marked the first occasion where a Chinese tech company was credited as the primary affiliation on a paper published in Nature.

The publication elucidates the process of using a well-established approach from the field of natural language processing (NLP) to tackle complex biological problems. It provides insight into how a seemingly simple solution, generally used to comprehend words and grammar, can be applied to this scenario.

Messenger RNA (mRNA) technology has emerged as a revolutionary force in vaccine development and the potential treatment of various diseases, including cancer. As a crucial messenger delivering genetic instructions from DNA to the cell's protein-producing machinery, mRNA plays a key role in generating specific proteins required for numerous functions within the human body. Due to its safety, efficacy, and ease of production, mRNA has been rapidly adopted in COVID-19 vaccine development.

Overview of mRNA coding region design for two well-established objectives, stability and codon optimality, using SARS-CoV-2 Spike protein as an example. (CREDIT: Baidu Research)

However, the inherent instability of mRNA poses significant challenges. This instability can result in inadequate protein expression, thereby weakening a vaccine's ability to induce potent immune responses. Additionally, it creates hurdles in storing and transporting mRNA vaccines, especially in developing countries with limited resources.

Previous studies have demonstrated that optimizing the secondary structure stability of mRNA, in conjunction with optimal codons, leads to improved protein expression. However, the issue lies in the vastness of the mRNA design space due to synonymous codons. There are approximately 10^632 mRNAs that can encode the same SARS-CoV-2 Spike protein, which makes prior methodologies practically unfeasible.

Experimental evaluation of LinearDesign-generated mRNA sequences encoding SARS-CoV-2 Spike protein. (CREDIT: Baidu Research)

Although NLP and biology may seem disparate at first glance, they share robust mathematical ties. Like a sentence in human language consisting of a word sequence and an underlying syntactic tree that conveys meaning, an RNA strand possesses a nucleotide sequence and a related secondary structure based on its folding pattern.

The team employed a technique in language processing known as lattice parsing, which represents potential word connections in a lattice graph and selects the most plausible option based on grammar rules. Similarly, they devised a graph that compactly represents all mRNA candidates, using deterministic finite-state automaton (DFA). In this context, finding the optimal mRNA is similar to identifying the most likely sentence among an array of similar alternatives.

The "wildtype" RNA sequence contains numerous unpaired nucleotide loops, which leads to less stability. The LinearDesign algorithm generates sequences with significantly fewer loops, a more stable structure. (CREDIT: Baidu Research)

The sophisticated yet straightforward approach of LinearDesign enabled the generation of the most stable mRNA sequence encoding the Spike protein in a remarkable 11 minutes.

In comparative studies, the sequences designed by LinearDesign outperformed existing vaccine sequences significantly. The algorithm achieved up to a 5-fold increase in stability (mRNA half-life), a 3-fold increase in protein expression levels (within 48 hours), and a staggering 128-fold increase in antibody response for COVID-19 mRNA vaccine sequences. For VZV mRNA vaccine sequences, the research reported up to a 6-fold increase in stability, a 5.3-fold increase in protein expression levels, and an 8-fold increase in antibody response.

Dr. Zhang further highlighted the implications of these findings. “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. This will greatly reduce the vaccine research and development costs for biopharmaceutical companies while improving the outcomes,” he stated.

In 2021, Baidu and pharmaceutical giant Sanofi embarked on a partnership to incorporate the LinearDesign algorithm into Sanofi's product design pipeline for mRNA vaccine and drug development.

Baidu has developed a bio-computing platform called PaddleHelix, based on PaddlePaddle. It encompasses the ERNIE-Bio-Computing Big Models and explores the application of AI in various domains, such as small molecules, proteins/peptides, and RNA.

This platform is pioneering a new research paradigm for AI in life sciences. Baidu's ERNIE Big Model has created a comprehensive big model technology system, including NLP, vision, cross-modal, and bio-computing. The recently introduced ERNIE Bot, a knowledge-enhanced large language model (LLM) capable of understanding and generating human language, is a part of the ERNIE Big Model family.

In the future, Baidu intends to further delve into AI applications in life sciences, broadening the range and depth of inclusive technology. By championing the health and well-being of all, the company is leading the charge in making a profound impact on humanity.


Note: Materials provided above by The Brighter Side of News. Content may be edited for style and length.

Like these kind of feel good stories? Get the Brighter Side of News' newsletter.


Joseph Shavit
Joseph ShavitSpace, Technology and Medical News Writer
Joseph Shavit is the head science news writer with a passion for communicating complex scientific discoveries to a broad audience. With a strong background in both science, business, product management, media leadership and entrepreneurship, Joseph possesses the unique ability to bridge the gap between business and technology, making intricate scientific concepts accessible and engaging to readers of all backgrounds.