论文部分内容阅读
前列腺癌病因及发病机理研究有助于前列腺癌预防和治疗.目前,前列腺癌生化试验研究方法成本高、耗时,而基于网络计算方法容易受基因表达谱数据不完整、噪声高及实验样本数量少等约束.为此,本文提出一种基于节点-模块置信度及局部模块度的双重约束算法(命名为NMCOM),挖掘前列腺癌候选疾病模块.NMCOM算法不依赖基因表达谱数据,采用候选基因与致病表型之间一致性得分,候选基因与致病基因之间语义相似性得分融合排序策略,选取起始节点,并基于节点-模块置信度及局部模块度双重约束挖掘前列腺癌候选疾病模块.通过对挖掘出的模块进行富集分析,最终得到18个有显著意义的候选疾病基因模块.与单一打分排序方法及随机游走重开始方法相比,NMCOM融合排序策略的平均排名比小、AUC值大,且挖掘出结果明显优于其他模块挖掘算法,模块生物学意义显著.NMCOM算法不仅能准确有效地挖掘前列腺癌候选疾病模块,且可扩展挖掘其他疾病候选模块.
Prostate cancer etiology and pathogenesis research contribute to the prevention and treatment of prostate cancer.Currently, prostate cancer biochemical test research method is costly and time-consuming, and network-based calculation method is vulnerable to incomplete gene expression data, high noise and the number of experimental samples Less constraint etc. Therefore, this paper proposes a dual constraint algorithm (named as NMCOM) based on the node-module confidence and local module degree to mine the candidate disease module of prostate cancer.NMCOM does not depend on the gene expression profile data and adopts the candidate gene And the pathogenic phenotype, the semantic similarity score between candidate gene and pathogenicity gene fusion sorting strategy, select the starting node, and based on the node-module confidence and local modular double-restriction mining prostate cancer candidate disease Finally, 18 gene modules with significant candidate disease genes were obtained by enrichment analysis of the excavated modules.Compared with the single scoring method and the random walk resumption method, the average ranking of the NMCOM fusion ranking strategy is small , AUC value is large, and the result of mining is obviously better than other module mining algorithms, the significance of module biology is significant.NMCOM Method can not only accurately and efficiently excavate module candidate diseases prostate cancer, and other diseases scalable Mining candidate modules.