AI framework AGENT yields thermostable solid-state mRNA–LNP vaccines in Nature Biotechnology study
A 28 September 2026 Nature Biotechnology research briefing summarises peer-reviewed work by Tian and colleagues on AGENT, an AI framework that combined high-throughput experiments with Bayesian optimization to design thermostable solid-state mRNA–lipid nanoparticle vaccines retaining full bioactivity after two months at 37 °C in preclinical animal studies.
A research briefing published in Nature Biotechnology on 28 September 2026 summarises a peer-reviewed paper by Tian and colleagues describing AGENT, an artificial-intelligence framework that combines high-throughput experiments with Bayesian optimization to design thermostable solid-state mRNA–lipid nanoparticle (LNP) vaccine formulations.
What the study reported
According to the briefing and the primary paper (DOI 10.1038/s41587-026-03331-w), optimized formulations retained full bioactivity after storage at 37 °C for two months. Immune responses in the reported experiments were non-inferior to those elicited by fresh vaccines. The authors also demonstrated cold-chain-free delivery via microneedle patches in rodents and nonhuman primates.
Study type and limits — read carefully
These are preclinical animal data. The work has not yet been tested in human clinical trials. The Nature Biotechnology research briefing summarises the peer-reviewed primary paper; it is not itself a clinical guideline. Results in animals often fail to translate directly to people. This article is not personal medical advice — talk to a doctor about vaccines or immunisation decisions.
Also available in: Саха тылаРусскийEspañol中文Portuguêsالعربية