Frontiers in cellular and infection microbiology · 2026
Vaxjo 2.0: An ontology- and large language model-powered knowledge base of vaccine adjuvants and mechanisms.
СтатьяНазвание и аннотация: Перевод в очереди
Обработка статьи · этапы, ошибки и управление
Загрузка состояния…
Ошибки и ограничения
История · было → стало
Состояние обновляется каждые 5 секунд, пока блок открыт. Процент показывается только там, где известен объём работы. Пауза останавливает продолжение этапа; текущий запрос к источнику может завершиться.
Monickaraj Joshua, Rehana Hasin, Bernardi Ani, Zheng Yuping, Lin Taiyu, Liu Le, Madireddi Amogh, Yeh Leo, Zheng Jie, Hur Junguk, He Yongqun
Аннотация
INTRODUCTION: Vaccine adjuvants enhance immune responses by boosting vaccine efficacy, reducing required doses, and improving long-term immunity. The Vaxjo database is a web-based resource that stores information on vaccine adjuvants, including their names, storage conditions, structures, preparation methods, components, functions, safety, and references. The original version of Vaxjo, released in 2012 with 103 vaccine adjuvants, has been expanded and modernized as Vaxjo 2.0, a significantly enhanced version.
METHODS: In Vaxjo 2.0, newly identified adjuvants from biomedical literature retrieved through PubMed searches, as well as from the Vaccine Adjuvant Compendium (VAC) and AdjuvareDB, were added. To accelerate data collection and curation, we developed a vaccine adjuvant large language model (Vaxjo-LLM) system that automatically identifies and annotates new vaccine adjuvants and characterizes their mechanisms. The LLM-mined results were manually evaluated, annotated, and selectively included in the database to ensure quality.
RESULTS: Overall, Vaxjo 2.0 includes 448 vaccine adjuvants, organized into 16 distinct categories (e.g., mineral salt, emulsion, cytokine, peptide, and toll-like receptor (TLR) agonist vaccine adjuvants), all of which are represented in the Vaccine Ontology (VO) to streamline information storage and exchange. From 817 PubMed abstracts, Vaxjo-LLM identified mechanisms for 323 unique vaccine adjuvants across 16 mechanism families, including T cell activation/polarization, dendritic cell activation, TLR signaling, inflammasome activation, cytokine signaling, B cell/antibody production, and pattern recognition receptor (PRR) sensing. The mined information was subsequently manually reviewed to ensure consistency and accuracy. Based on this analysis, the mechanism profiles and clustering of the top 20 adjuvants were generated, revealing shared and distinct mechanistic signatures. For deeper mechanistic understanding, Vaxjo 2.0 further classified adjuvants based on PRR families, including TLRs, C-type lectin receptors, NOD-like receptors, and RIG-I-like receptors. Vaccine adjuvants were also categorized based on host immune response profiles, such as Th1/Th2/Th17-biased, Th1/Th2-mixed, and Treg-biased immune profiles.
DISCUSSION: The newly designed Vaxjo 2.0 web interface (https://violinet.org/vaxjo) provides an openly available and user-friendly platform for querying, visualizing, and analyzing vaccine adjuvant data.
METHODS: In Vaxjo 2.0, newly identified adjuvants from biomedical literature retrieved through PubMed searches, as well as from the Vaccine Adjuvant Compendium (VAC) and AdjuvareDB, were added. To accelerate data collection and curation, we developed a vaccine adjuvant large language model (Vaxjo-LLM) system that automatically identifies and annotates new vaccine adjuvants and characterizes their mechanisms. The LLM-mined results were manually evaluated, annotated, and selectively included in the database to ensure quality.
RESULTS: Overall, Vaxjo 2.0 includes 448 vaccine adjuvants, organized into 16 distinct categories (e.g., mineral salt, emulsion, cytokine, peptide, and toll-like receptor (TLR) agonist vaccine adjuvants), all of which are represented in the Vaccine Ontology (VO) to streamline information storage and exchange. From 817 PubMed abstracts, Vaxjo-LLM identified mechanisms for 323 unique vaccine adjuvants across 16 mechanism families, including T cell activation/polarization, dendritic cell activation, TLR signaling, inflammasome activation, cytokine signaling, B cell/antibody production, and pattern recognition receptor (PRR) sensing. The mined information was subsequently manually reviewed to ensure consistency and accuracy. Based on this analysis, the mechanism profiles and clustering of the top 20 adjuvants were generated, revealing shared and distinct mechanistic signatures. For deeper mechanistic understanding, Vaxjo 2.0 further classified adjuvants based on PRR families, including TLRs, C-type lectin receptors, NOD-like receptors, and RIG-I-like receptors. Vaccine adjuvants were also categorized based on host immune response profiles, such as Th1/Th2/Th17-biased, Th1/Th2-mixed, and Treg-biased immune profiles.
DISCUSSION: The newly designed Vaxjo 2.0 web interface (https://violinet.org/vaxjo) provides an openly available and user-friendly platform for querying, visualizing, and analyzing vaccine adjuvant data.