FIS-PNN: A hybrid computational method for protein-protein interaction prediction

The study of protein-protein interactions (PPI) is an active area of research in biology as it mediates most of the biological functions in any organism. Although, there are no concrete properties in predicting PPI, extensive wet-lab experiments suggest (with a high probability) that interacting proteins in the fine level share similar functions, cellular roles and sub-cellular locations. In this study, we developed a technique to predict PPI based on their secondary structures, co-localization, and function annotation. We proposed our approach, namely FIS-PNN, to predict the interacting proteins in yeast using hybrid machine learning algorithms. FIS-PNN has been trained and tested using 1029 proteins with 2965 known positive interactions; it could successfully predict PPI with 96% of accuracy - a level that is significantly greater than all other existing sequence-based prediction methods.