Vaxjo 2.0 Supplementary Files and Source codes

This dataset comprises data related to the Vaxjo database, including: Supplementary File 1 - Supp1_Vaxjo-LLM-2Phase-End2EndPipeline.ipynb: Complete code used to implement the two-phase LLM workflow for extracting and summarizing vaccine adjuvant mechanisms from PubMed abstracts. The file includes the full extraction module and the secondary summarization module used in Vaxjo 2.0. Supplementary File 2 - Supp2_vaccine_adjuvants.xlsx: The master Excel file contains a comprehensive table of all 448 curated vaccine adjuvants in Vaxjo, including the adjuvant source, Vaxjo record link, Vaxjo name, VO ID, VO label, adjuvant category, target receptor(s), and induced immune profile(s) (if available). Supplementary File 3 - Supp3_Vaxjo-LLM-Stage1-Response.txt: LLM prompt 1 and output 1 for getting adjuvant mechanisms . Contains the exact prompt used in Phase I to extract immune-response mechanism text for each adjuvant, together with representative raw model outputs before expert curation. Supplementary File 4 - Supp4_Vaxjo-LLM-Stage2-Response.txt: LLM prompt 2 and output 2 file for summarizing the adjuvant mechanisms . Contains the summarization prompt used in Phase II to generate concise mechanism summaries and mechanism-family assignments, together with representative model outputs prior to expert review. Table1_DL_queries.txt: The DL queries used to generate the counts listed in Table 1. VO (folder): The source code for the Vaccine Ontology (version 2025-12-03) was cloned from the GitHub repository (https://github.com/vaccineontology/VO) on December 3, 2025. Vaxjo-LLM (folder): The source code for Vaxjo-LLM, along with input and output files, was cloned from the GitHub repository (https://github.com/hurlab/Vaxjo-LLM) on December 4, 2025.