Figure 1.

Schema of the integrated and semi-automated curation workflow. The curation process starts with the automatic search and retrieval of PubMed records. Text mining methods support the annotation of relevant entities (antimicrobial agents, methods for antimicrobial susceptibility testing, type of agent combination, and organisms) and the assessment of document relevance. Articles deemed relevant are further curated manually by experts that revise automatic annotations and look into additional information about the types of combinations and other relevant details of the experiments. Articles describing previously uncharacterized AMP combinations are added to the network.