Dynamic Complex Protein Detection using Binary Harris Hawks Optimization

Identifying protein complexes brings new insights in the field of understanding of cellular life and the mechanisms underlying complex diseases. In order to solve the challenge of protein complex prediction, a large number of tools have been developed to create an efficient approach. Most existing methods do not take changing protein-protein interactions with time into consideration. To address these challenges, we propose a new binary technique of Harris Hawks Optimizer (BHHO) to improve the accuracy of protein complex detection. According to the core-attachment structure, BHHO is used to discover the potential cores of protein complexes by simulating the process of cooperative behavior in chasing by Harris' hawks in nature. We design a new fitness function to detect protein complexes with various densities and modularities. The experimental results show that BHHO achieves an ideal performance PIN in term of the F1 and outperforms of the classical algorithms in yeast in term Jaccard measure.