利用蛋白质-表型网络的致病基因预测方法研究

在“表型相似的疾病是由功能相关的基因引起”这一假设基础上，提出了一种利用人类蛋白质相互作用和疾病表型相似性网络进行疾病候选基因预测的新方法，同时开发了候选基因预测系统——GENEDIG，该方法通过建立回归模型，利用向量相关性分析，对诱发基因未知疾病的每一个候选基因计算得分，并根据得分结果进行排序，达到预测致病基因的目的。该方法还可进一步用于探讨多个基因在复杂疾病中的协同性。实验结果表明，该方法能有效揭示疾病与诱发基因之间的联系，为进一步的生物学验证实验提供帮助。

Prediction of human disease genes based on protein-phenotype network

On the basis of the assumption that diseases with similar phenotypes are caused by functionally related genes,a candidate gene prioritization method that is entirely based on the network of human protein-protein interactions and disease phenotypic similarities is described.A tool named GENEDIG is developed.A regression model is established and a vector relativity analysis method is applied to prioritize all candidate genes according to the probability of causing certain disease phenotypes.This method is applicable to genetically uncharacterized phenotypes,and also extendable to explore gene cooperativity in complex diseases.This method can effectively reveal the implied association between disease and induced genes to facilitate future discovery of disease genes.