Detecting protein complexes from DPINs by density based clustering with Pigeon-Inspired Optimization Algorithm

Abstract

Detecting protein complexes is crucial to understand principles of cellular organization. Plenty evidences have indicated that sub-graphs with high density in protein-protein interaction (PPI) network, especially dynamic PPI network (DPIN), usually correspond to protein complexes. As a well-known density-based clustering algorithm, Density-Based Spatial Clustering of Applications with Noise (DBSCAN) has been used in many areas due to its simplicity and the ability to detect clusters of different sizes and shapes. However, one of its limitations is that the performance of DBSCAN depends on two specified parameters $\varepsilon$ and MinPts, where $\varepsilon$ represents the maximum radius of a neighborhood from an observing point while MinPts means the minimum number of data points contained in such a neighborhood. In this article, we develop a new method named as P-DBSCAN to detect protein complexes in DPIN by using Pigeon-Inspired Optimization (PIO) Algorithm to optimize the parameters $\varepsilon$ and MinPts in DBSCAN. The experiments on DIP and MIPS datasets show that P-DBSCAN outperforms the state-of-the-art methods for protein complex detection in terms of several criteria such as precision, recall and f-measure.