Figure 1

Data-mining pipeline. The process of mining PubMed abstracts can be divided into 4 main steps: 1) Finding age-related abstracts, 2) determining the age-phenotype relationship type for those age-related abstracts, 3) generating a textual snippet which describes the most important information given in the abstract regarding the captured age and 4) mapping the text snippets to phenotypes from the DO and a subset of the UMLS. Instances for which an age-phenotype relationship could not be determined or no phenotype were found were not used to populate the database.