Premium
Construction and application of dynamic protein interaction network based on time course gene expression data
Author(s) -
Wang Jianxin,
Peng Xiaoqing,
Li Min,
Pan Yi
Publication year - 2013
Publication title -
proteomics
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.26
H-Index - 167
eISSN - 1615-9861
pISSN - 1615-9853
DOI - 10.1002/pmic.201200277
Subject(s) - sigma , computer science , expression (computer science) , time point , computational biology , matching (statistics) , gene expression , gene , biology , biological system , genetics , mathematics , physics , statistics , quantum mechanics , acoustics , programming language
In recent years, researchers have tried to inject dynamic information into static protein interaction networks ( PIN s). The paper first proposes a three‐sigma method to identify active time points of each protein in a cellular cycle, where three‐sigma principle is used to compute an active threshold for each gene according to the characteristics of its expression curve. Then a dynamic protein interaction network ( DPIN ) is constructed, which includes the dynamic changes of protein interactions. To validate the efficiency of DPIN , MCL , CPM , and core attachment algorithms are applied on two different DPIN s, the static PIN and the time course PIN ( TC ‐ PIN ) to detect protein complexes. The performance of each algorithm on DPIN s outperforms those on other networks in terms of matching with known complexes, sensitivity, specificity, f ‐measure, and accuracy. Furthermore, the statistics of three‐sigma principle show that 23–45% proteins are active at a time point and most proteins are active in about half of cellular cycle. In addition, we find 94% essential proteins are in the group of proteins that are active at equal or great than 12 timepoints of GSE 4987, which indicates the potential existence of feedback mechanisms that can stabilize the expression level of essential proteins and might provide a new insight for predicting essential proteins from dynamic protein networks.