< My PhD

The VaccO ontology of vaccine descriptions

A: Code alignment

Applications: https://app.vac4eu.org/vacco

Code: https://github.com/mi-erasmusmc/vacco (public after publication)


Alignment of Vaccine Codes Using the VaccO Ontology of Vaccine Properties.
Benedikt Becker, Jan Kors, Erik Mulligen, Miriam Sturkenboom.
Submitted, 2018.


Motivation Vaccine information is represented in European electronic health record (EHR) databases using various clinical and database-specific coding systems and drug vocabularies. The lack of harmonization constitutes a challenge in reusing EHR data in collaborative benefit-risk studies about vaccines.

Results We designed an ontology of vaccine properties, VaccO, and implemented algorithms for the analysis of vaccine code descriptors and for the alignment of vaccine coding systems, i.e., the identification of corresponding codes from different coding systems. The alignment algorithm demonstrated excellent performance in a comparison with two manually created reference sets including clinical and database-specific coding systems using multilingual descriptors (F-scores 0.91 and 0.96). The automatic alignment of vaccine coding systems accelerates the readiness of EHR databases in collaborative vaccine studies.


Figure 1: Structure of the core VaccO ontology.
Figure 2: Example for the compilation of vaccine code descriptors in VaccO. VaccO classes are identified in the code descriptors (blue boxes in the source and target code descriptors) and compiled into vaccine classes (green boxes in the center), which are defined in terms of the VaccO classes (yellow boxes on the right).

B: Vaccine identification


  • A gold standard of 150 abstracts (and titles) about human vaccines with NER annotations of vaccine properties
  • A baseline for NER of vaccine properties using the VaccO ontology
  • Comparison of ontology-based and deep learning approaches for vaccine normalization of scientific literature (assignment of MeSH vaccine codes)


Identification and normalization of vaccine descriptions in scientific literature.
Benedikt Becker, Miriam Sturkenboom, Jan Kors.
Submitted, 2018.

Last update: 2021-06-30