融合基因芯片数据上下文特定网络的重构

目的在特定类型细胞或特定环境条件下，已有的全基因组代谢网络中的部分反应并不参与到实际的代谢过程中。为反映特定条件下生物体代谢系统的工作状态，本文提出一种重构上下文特定网络的新方法，用来构建特定条件下的代谢网络。方法首先根据基因芯片数据中蕴含的PACalls信息，南基因-酶-反应之间的调控关系，计算出在表达层面上为“Absent”状态的反应集合。然后使用一个？昆合整数规划模型，从全基因代谢网络中删除表达为“Absent”状态的反应及其关联反应，在满足计量平衡等约束条件下．寻找与表达层面吻合最优的网络，即为特定条件下的上下文特定网络。结果重构了肝脏和心脏两种组织的上下文特定网络。实验结果表明，两种组织的上下文特定网络存在很大差异性．尤其在某些具有组织特异性的代谢功能上，如胆汁酸合成等。结论融合基因芯片数据信息的上下文特定网络重构方法，能够有效构建反映特定条件下生物体代谢系统状态的上下文特定网络。

Context-specific network reconstruction of integrating microarray data

Objective Some reactions in a genome-scale network are inactive in a specific cell type or under a particularly environmental condition. In order to characterize the current state of metabolic system of organism,we propose a new method to reconstruct the context-specific network. Methods The probeset detection information （PA Calls） of microarray data was adopted to infer ＇ Absent＇ reactions by gene-protein-reaction relationships and a mixed integer linear programming model was applied to remove ＇Absent＇ reactions and cascaded reactions from the genome-scale network directly. The context-specific network thus generated satisfied stoichiometric constraints and directional constraints. Results Applying this method to human metabolism, we reconstructed the context-specific networks of normal heart and liver. The results showed that the context- specific networks of the two tissues were significantly different, especially on some tissue-specific metabolic functions such as bile acid biosynthesis. Conclusions By integrating microarray data,this method reconstructed context-specific network that reflected the state of metabolic system under a particular condition.