融合PPI网络和基因表达的复合物识别算法

从大规模相互作用网络中识别蛋白质复合物，对解释特定的生物进程和预测蛋白质功能具有重要作用，同时也是后基因组时代一 个最重要的研究课题. 考虑到传统仅基于蛋白质相互作用网络（PPI网络）的蛋白质复合物识别算法可靠性不高，本文提出 了一种新的融合PPI网络和基因表达数据的蛋白质复合物识别算法IPCIPG. 区别于之前用基因表达数据评估PPI网络可靠性的做法，本文提出在蛋白质复合物的识别过程中将PPI网络和基因表达数据有机地结合起来. 算法IPCIPG首先根据边聚集系数（ECC）与蛋 白质间共表达的相关性（PCC）计算PPI网络中每个节点的权重，权重最大的节点作为种子，然后从种子节点开始扩充生成稠密子图. 基于酵母数据集的实验结果表明，算法IPCIPG较其他算法HUNTER，HC-PIN，CMC，SPICI，MOCDE，MCL能够更准确，更有效地 识别出具有特定生物意义的蛋白质复合物.

An algorithm for identifying protein complexes based on the integration of PPI network and Gene expression

Identifying protein complexes from the large-scale protein interaction network is crucial to understand principles of cellular organization and predict protein functions, which is one of the most important issues in the post-genomic era. Generally, the traditional protein complex discovery algorithms are only based on the protein-protein interaction network (PPI network), and are not so accurate. In this paper, a novel algorithm IPCIPG is proposed based on the integration of the PPI network and the gene expression data. Different from other previous methods which use gene expression data to evaluate the reliability of PPIs, IPCIPG integrates the gene expression data into PPI network during the identification of protein complexes. IPCIPG uses the edge clustering coefficient (ECC) and the co-expression correlation between proteins (PCC) to calculate the weight of each node in the PPI network. And then the node with the highest weight is selected as seed, then, a dense sub-graph will be obtained by extending from the seed. The experiment results on the data of Saccharomyces cerevisiae show that IPCIPG can identify the protein complexes with specific biological meaning more effectively, precisely and comprehensively than the other algorithms HUNTER, HC-PIN, CMC, SPICI, MOCDE, and MCL.