Maksim Goncharenok / Pexels · Pexels License · Image source
Research: what does the algorithm improve?
Published in Nature in 2023, LinearDesign uses an algorithm from computational linguistics to optimize mRNA structural stability and codon usage together. This differs from a conversational model. Researchers compared COVID-19 and varicella-zoster vaccine candidates with standard codon optimization, evaluating mRNA stability, protein production and immune responses in mice. These experiments alone do not establish protection against disease in humans or authorization of a product.
Clinical evidence: human studies remain essential
WHO describes preclinical evaluation followed by three phases of human trials. Initial assessments of safety, dose and immune responses lead to larger studies of efficacy and safety. Authorities then review the evidence, and monitoring continues after introduction. A design score, a laboratory measurement and a clinical trial outcome should therefore not be treated as equivalent levels of evidence.
Regulation: quality needs its own assessment
When checked on 7 October 2026, EMA’s page still listed its mRNA vaccine quality document as a draft with consultation closed. It addresses manufacturing, characterization, specifications and analytical controls for mRNA vaccines against infectious diseases. Draft guidance is not a marketing authorization for any particular vaccine; we do not present it here as an adopted final guideline.
Editorial perspective: five questions for an AI vaccine story
Our reading framework is to ask five separate questions. What task does the algorithm perform? What is the comparator? Is the measured outcome a molecular property, an animal finding or a clinical outcome in humans? What manufacturing and quality evidence is available? Can the specific product’s regulatory status be verified in the relevant country? Keeping these answers separate helps readers avoid placing promising design research at a more advanced clinical stage than the evidence supports. This is a framework for reading research, not a vaccine recommendation or an efficacy claim for RAI products.
Sources
- Zhang et al. — Algorithm for optimized mRNA design improves stability and immunogenicity. Nature (2023) ↗
Accessed: 2026-10-07 - WHO — How are vaccines developed? ↗
Accessed: 2026-10-07 - EMA — Development of a guideline on the quality aspects of mRNA vaccines (EMA/CHMP/BWP/82416/2025; draft) ↗
Accessed: 2026-10-07
This article is for education and information. Referencing an organization does not imply its endorsement of RAI PACS Pro. This is not a claim of clinical effectiveness for a product or model.