A novel extended Pareto Optimality Consensus model for predicting essential proteins. (7th November 2019)

• Record Type: Journal Article
• Authors: Li, Gaoshi; Li, Min; Peng, Wei; Li, Yaohang; Pan, Yi; Wang, Jianxin
• Is Part Of: Journal of theoretical biology. Volume 480(2019)
• Page Start: 141
• Page End: 149
• Publication Date: 2019-11-07
• Subjects: Essential proteins -- PPI network -- Orthologous -- Gene expression; Biology -- Periodicals; Biological Science Disciplines -- Periodicals; Biology -- Periodicals; Biologie -- Périodiques; Theoretische biologie; Biology; Periodicals; 571.05
• DOI: 10.1016/j.jtbi.2019.08.005
• Languages: English
• ISSNs: 0022-5193

Abstract

Highlights: A novel method for identifying essential proteins named as E_POC is proposed. A now combination model based on POC is developed to fuse two types of information. E_POC outperforms the existing classical methods on S. cerevisiae and E. coli data. Abstract: Essential proteins have vital functions, when they are destroyed in cells, the cells will die or stop reproducing. Therefore, it is very important to identify essential proteins from a large number of other proteins. Due to the time-consuming, expensive, and inefficient process in biological experimental methods, computational methods become more and more popular to recognize them. In the early stages, these methods mainly rely on protein-protein interaction (PPI) information, which limits their discovery capacities. Researchers find novel methods by fusing multi-information to improve prediction accuracy. According to these features, essential proteins are more important and conservative in the evolution process, their neighbors in PPI networks are usually likely to be essential, there are many false positives in PPI data, whether a protein is essential can be assessed by the importance of a protein itself, the relevance of neighbors and the reliability of PPIs. The importance of neighbors and the reliability of PPIs can be further integrated into neighborhood feature. In the study, orthologous information, edge-clustering coefficient and gene expression information are used to measure the importance of a protein itself, the importance of the neighbors and the reliability of PPIs, respectively. Then, a novel expanded POC model, E_POC, is proposed to fuse the above information to discover essential proteins, a weighted PPI network is constructed. The proteins ranked high according to their weights are treated as candidate essential proteins. This novel method is named as E_POC. E_POC outperforms the existing classical methods on S. cerevisiae and E. coli data. Graphical abstract: Whether proteins are essential or not is determined by two factors: whether themselves are important and whether their neighbors are important. Therefore, these two factors are scored separately, and then uses the E_POC model to fuse the two scores into one score, the proteins with high score are essential proteins candidates.Image, graphical abstract