论文部分内容阅读
肺癌致病基因的发现及预测有助于认识肺癌的发生机理、诊断与防治,是人类基因组研究的重要目标。应用现有二元网络重启随机游走算法预测致病基因时,一般先在疾病表型网络、蛋白质作用网络及疾病-蛋白质二分图网络内随机游走一步,然后进行网络间跳转,这种策略不仅搜索效率较低,还可能遗漏蛋白质(或疾病)网络中的局部拓扑信息。鉴于此,作者提出一种二元网络异步重启游走(asynchronously random walk with restart,ARWRH)算法,构建疾病表型-蛋白质异构网络,深层次挖掘潜在肺癌风险致病基因。ARWRH算法首先在疾病表型网络、蛋白质作用网络及疾病表型-蛋白质二分图网络内随机游走不同步数,然后进行网络间跳转,迭代形成稳态概率向量,从而获得候选致病基因。仿真实验表明,ARWRH算法可有效预测肺癌潜在风险致病基因,多数预测结果获得了文献证据支持。
The discovery and prediction of pathogenic genes in lung cancer help to understand the pathogenesis, diagnosis and prevention of lung cancer, which is an important goal of human genome research. When using the existing binary network restart random walk algorithm to predict pathogenic genes, it usually first walks one step randomly in disease phenotype network, protein action network and disease-protein bipartite graph network, then jumps between networks Strategies not only search less efficiently but may also miss local topology information in protein (or disease) networks. In view of this, the authors propose a binary network asynchronously random walk with restart (ARWRH) algorithm to construct a disease phenotype - protein isomerism network, dig deep potential risk genes for lung cancer. The ARWRH algorithm randomly walks out-of-sync numbers in disease phenotype networks, protein action networks, and disease phenotype-protein bipartite networks, and then performs inter-network jumps to form steady-state probability vectors iteratively to obtain candidate pathogenic genes. Simulation results show that the ARWRH algorithm can effectively predict the potential risk of lung cancer genes, the majority of prediction results obtained the literature support.