一种基于复合物参与度的关键蛋白质预测方法

A Method for Predicting Essential Protenis Based on Degree of Participation in Protein Complexes

针对在交互网络上的关键蛋白质识别通常只关注节点在拓扑层面上的一些特性,但关键蛋白质识别过程中有必要结合蛋白质功能方面的信息,提出了一种基于复合物参与度的关键蛋白质预测方法 DPC(Degree of Participation in Complexes),该方法考虑节点在复合物中邻居节点的数量,综合了蛋白质在复合物内部以及在全局总体上的拓扑特性.在酵母蛋白质网络上的实验结果表明,DPC在关键蛋白质识别率方面明显优于其他六种经典的节点中心性拓扑参数,并且DPC能够识别出参与多个复合物的蛋白质,与关键蛋白质的生物意义相吻合.

Considering the identification of essential proteins in the interaction network tended to focus only on topological characteristics of the nodes, however, it is necessary to combine information about protein function to identify the essential proteins. A new method for predicting essential proteins based on degree of participation in protein complexes named DPC（Degree of Participation in Complexes） is proposed. The method considers the number of neighbors inside protein complexes, the topological characteristics of nodes both inside their protein complexes and in the whole network is considered. The experimental results in yeast protein interaction networks show the number of essential proteins discovered by DPC universally exceeds that discovered by other six classical centrality measures. Moreover, DPC can identify proteins which participate in multiple complexes. This result consistent with the biological significances of essential proteins.