The immune epitope database: a historical retrospective of the first decade

Immunology. 2012 Oct;137(2):117-23. doi: 10.1111/j.1365-2567.2012.03611.x.


As the amount of biomedical information available in the literature continues to increase, databases that aggregate this information continue to grow in importance and scope. The population of databases can occur either through fully automated text mining approaches or through manual curation by human subject experts. We here report our experiences in populating the National Institute of Allergy and Infectious Diseases sponsored Immune Epitope Database and Analysis Resource (IEDB,, which was created in 2003, and as of 2012 captures the epitope information from approximately 99% of all papers published to date that describe immune epitopes (with the exception of cancer and HIV data). This was achieved using a hybrid model based on automated document categorization and extensive human expert involvement. This task required automated scanning of over 22 million PubMed abstracts followed by classification and curation of over 13 000 references, including over 7000 infectious disease-related manuscripts, over 1000 allergy-related manuscripts, roughly 4000 related to autoimmunity, and 1000 transplant/alloantigen-related manuscripts. The IEDB curation involves an unprecedented level of detail, capturing for each paper the actual experiments performed for each different epitope structure. Key to enabling this process was the extensive use of ontologies to ensure rigorous and consistent data representation as well as interoperability with other bioinformatics resources, including the Protein Data Bank, Chemical Entities of Biological Interest, and the NIAID Bioinformatics Resource Centers. A growing fraction of the IEDB data derives from direct submissions by research groups engaged in epitope discovery, and is being facilitated by the implementation of novel data submission tools. The present explosion of information contained in biological databases demands effective query and display capabilities to optimize the user experience. Accordingly, the development of original ways to query the database, on the basis of ontologically driven hierarchical trees, and display of epitope data in aggregate in a biologically intuitive yet rigorous fashion is now at the forefront of the IEDB efforts. We also highlight advances made in the realm of epitope analysis and predictive tools available in the IEDB.

MeSH terms

  • Bibliometrics
  • Databases, Genetic
  • Epitopes / immunology*
  • Humans
  • Internet
  • Time Factors


  • Epitopes