Systematic Differences in Signal Emitting and Receiving Revealed by PageRank Analysis of a Human Protein Interactome

Most protein PageRank studies do not use signal flow direction information in protein interactions because this information was not readily available in large protein databases until recently. Therefore, four questions have yet to be answered: A) What is the general difference between signal emitting and receiving in a protein interactome? B) Which proteins are among the top ranked in directional ranking? C) Are high ranked proteins more evolutionarily conserved than low ranked ones? D) Do proteins with similar ranking tend to have similar subcellular locations? In this study, we address these questions using the forward, reverse, and non-directional PageRank approaches to rank an information-directional network of human proteins and study their evolutionary conservation. The forward ranking gives credit to information receivers, reverse ranking to information emitters, and non-directional ranking mainly to the number of interactions. The protein lists generated by the forward and non-directional rankings are highly correlated, but those by the reverse and non-directional rankings are not. The results suggest that the signal emitting/receiving system is characterized by key-emittings and relatively even receivings in the human protein interactome. Signaling pathway proteins are frequent in top ranked ones. Eight proteins are both informational top emitters and top receivers. Top ranked proteins, except a few species-related novel-function ones, are evolutionarily well conserved. Protein-subunit ranking position reflects subunit function. These results demonstrate the usefulness of different PageRank approaches in characterizing protein networks and provide insights to protein interaction in the cell.

既往绝大多数蛋白质PageRank（protein PageRank）相关研究均未采用蛋白质相互作用中的信号流方向信息，直至近年，大型蛋白质数据库中才陆续收录此类方向数据。因此尚有四个核心问题亟待解答：A) 蛋白质相互作用组（protein interactome）内的信号发送与信号接收存在哪些一般性差异？B) 定向排名中位列前茅的蛋白质包含哪些类型？C) 排名靠前的蛋白质是否较排名靠后者具有更高的进化保守性？D) 排名相近的蛋白质是否倾向于拥有相似的亚细胞定位？本研究针对人类蛋白质信息定向网络，分别采用正向、反向与无向PageRank算法开展排名分析，并对其进化保守性进行探究，以此解答上述问题。正向排名将权重赋予信息接收方，反向排名将权重赋予信息发送方，而无向排名则主要依据相互作用数量进行计分。正向排名与无向排名生成的蛋白质列表呈现高度相关性，而反向排名与无向排名的蛋白质列表相关性则较低。研究结果显示，人类蛋白质相互作用组中的信号发送/接收系统具有「发送方关键化、接收方平均化」的特征。信号通路相关蛋白质常跻身排名前列。共有8种蛋白质同时位列信息发送与接收双榜的前列。除少数与物种特异性相关的新型功能蛋白质外，排名靠前的蛋白质均具有良好的进化保守性。蛋白质亚基的排名位次可反映其亚基功能。本研究结果证实了不同PageRank算法在表征蛋白质网络特征方面的应用价值，并为解析细胞内蛋白质相互作用机制提供了全新的认知视角。