《生命科学》 2008, 20(3):

评估蛋白质相互作用可信度的生物信息学方法

欧阳玉梅1,2,方若森2,马志强1*

(1东北师范大学计算机学院，长春130117；2伊犁师范学院，奎屯 833200)

摘　要：摘　要：随着基因组规模的高通量实验鉴定技术和计算预测方法的发展，出现了大量蛋白质相互作用数据，但大规模蛋白质相互作用数据中的较高比例的假阳性影响了相互作用数据的质量。生物信息学方法能够从已有的数据和知识出发，通过计算方法系统评估大规模蛋白质相互作用的可信度。本文从过程模型设计、数据集构建、特征选择与综合属性抽取、一些算法使用、实例概述等方面介绍了生物信息学方法评估蛋白质相互作用可信度的研究特点与进展。
关键词：蛋白质相互作用; PPI可信度; 生物信息学
中图分类号：TQ937; TP391　　文献标识码：A

Bioinformatics methods for assessment of the reliability of protein interactions

OUYANG Yu-mei1,2, FANG Ruo-sen2, MA Zhi-qiang1*

(1 College of Computer, Northeast Normal University, Changchun 130117, China; 2 Department of Computer, Yili Normal College, Kuitun 833200, China)

Abstract: Abstract: Large amounts of protein-protein interaction data have been produced with the development of various genome-scale high throughput experimental screening techniques and computational prediction approaches. As high throughput datasets are prone to higher false positive rates, it affects the expense of data quality. Bioinformatics methods assess the reliability of protein interactions from known data and knowledge by using computational methods. This paper introduces the characteristics and advances of bioinformatics methods for assessing the reliability of protein interactions by different aspects such as designing a process model, building the datasets, selecting characteristics, using some algorithms and describing some examples.
Key words: protein interactions; the reliability of PPI; bioinformatics