蛋白质交互网络模块化的数据分析

生物网络的模块化是生物系统鲁棒性的一个基本前提。首先根据已有的算法分析改进了经典的谱方法，提出了一个提高模块化方法分辨率的策略。采用不同来源的蛋白质网络来计算地分析网络的模块化属性，发现尽管这些蛋白质网络都展现了较高的模块化程度，但模块化的程度受到数据来源的影响。同时还发现基因在全基因组复制以后，会倾向于在同一个模块中出现，这说明复制基因功能不会产生特别大的差异，使得当删除一个复制基因时，模块内的另一个复制基因会有功能的补偿。

Data analysis of protein interaction network modularization

Modularization of biological network is a basic premise of biological systems robustness.First of all,this paper improves the classic method of spectrum based on the analysis of existing algorithms, and brings forward a strategy to enhance resolution of the modular approach.This paper uses different sources of protein network analysis to calculate property of network modularization,and finds the extent of the modularization is influenced by data source,despite these protein networks show a high degree of modularization.This paper also finds that the gene tends to be in the same module after the genome-wide copy,which shows that gene copy doesn＇t produce large differences in particular.When a gene copy is deleted, another gene copy should do the compensation function in the module.