Fig. 2

The workflow of DisPheno. It mainly contains four parts: a Annotating phenotype ontology information content using both gene annotation and disease annotation; b Reconstructing topological structure of phenotype term by calculating phenotype term definition similarity using TF-IDF; c Measuring phenotype semantic similarity based on HPO by integrating term definition-similarity; d Calculating phenotype term association and set similarity by measuring phenotype term associations using Point-wise Mutual Information(PMI)