Supervised Biomedical Semantic Similarity

Semantic similarity between concepts in knowledge graphs is essential for several bioinformatics applications, including the prediction of protein-protein interactions and discovery associations diseases genes. Although describe entities terms perspectives (or semantic aspects), state-of-the-art measures are general-purpose. This can represent a challenge since different use cases application may need ultimately depend on expert manual fine-tuning. We present new approach that uses supervised machine learning to tailor aspect-oriented fit particular view biological or relatedness. implement evaluate it using combinations representative methods with four views: interaction, protein function similarity, sequence phenotype-based gene similarity. The results demonstrate our outperforms non-supervised methods, producing models significantly better than commonly used aspects.