Abstract: The computational methods used to predict protein complexes from the protein interaction network have a great error because of the high false positive rate and false negative rate of protein interaction data. To compensate for this, a new protein complex prediction approach is proposed via the integration of multiply data sources. Matching analysis and GO functional enrichment analysis are performed as so to estimate the performance of the algorithm. The results show that the new algorithm is much better than previous ones.

Key words: protein interaction, gene expression profiles, essential protein, protein complex, matching statistics, gene ontology, functional enrichment

摘要： 蛋白质相互作用数据具有较高的假阳性率和假阴性率，这直接导致计算方法从中预测蛋白质复合物会产生较大的误差。为了弥补数据的这种先天性不足，通过结合多数据源，一种新的蛋白质复合物预测算法被提出。匹配分析和GO功能富集分析被用于评估算法的性能。测试结果表明，新算法远优于以前的其他算法。

关键词: 蛋白质相互作用, 基因表达谱, 关键蛋白质, 蛋白质复合物, 匹配统计, 基因本体, 功能富集