Predicting shrimp protein-protein interactions and gene ontology terms using association rule and semantic similarity calculation

Protein-protein interactions (PPIs) play a crucial part in cell functions. Detecting protein interactions by means of biology experiments are costly and very time consuming. Interactions of these proteins will result in different biochemical processes that govern living cells. Scientists have characterized gene functions using ontological structures in three categories, namely cellular component, biological process, and molecular function. Since genes code for proteins, establishing relationships between protein interactions and their refated ontologfcal definition would help predict unknown PPIs. In this paper, we present a novel method to predict new PPIs using information from proteins annotated in Gene Ontology. Our work can also predict Gene Ontology based on the association rules and the semantic similarity of Gene Ontology annotated to the proteins. We applied the proposed method to predict new PPIs using shrimp protein-protein interactions data and found that Gene Ontology and interactions between proteins can be used as a key to predict unknown protein-protein interactions. Moreover" we found the closely relationships between categories in Gene Ontology which are beneficial for protein-protein interaction prediction as well as for predicting other gene ontologies.