Fig. 1

The components and workflow of pyMeSHSim. pyMeSHSim consists of three subpackages, including metamapWrap, data and Sim. In bio-NE recognition, metamapWarp curates the UMLS concepts from free text. In bio-NE normalization, data translates UMLS concepts to MeSH terms, and maps SCRs to MHs using selected records and relationships between records in MeSH. In bio-NEs comparison, Sim uses IC-based and graph-based methods to measure semantic similarity between two bio-NEs