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

发布时间：2018-01-05 13:15

本文关键词：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　出处：《生物化学与生物物理进展》2015年04期 　论文类型：期刊论文

【摘要】：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.
[Abstract]:Studies on the etiology and pathogenesis of prostate cancer contribute to the prevention and treatment of prostate cancer. At present, the research methods of prostate cancer biochemistry test are expensive and time-consuming. However, the network-based computing method is prone to be constrained by incomplete data of gene expression spectrum, high noise and small number of experimental samples. This paper presents a dual constraint algorithm (named NMCOMs) based on node-module confidence and local modularity. Mining prostate cancer candidate disease module. NMCOM algorithm does not rely on gene expression profile data, using candidate gene and pathogenic phenotypic consistency score. The semantic similarity score between candidate gene and pathogenetic gene is fused to sort strategy and start node is selected. And based on the node-module confidence degree and local modular degree dual constraints mining prostate cancer candidate disease module. Through the enrichment analysis of the extracted module. Finally, 18 candidate gene modules with significant significance were obtained. Compared with single ranking method and random walk weight starting method, the average ranking of NMCOM fusion sequencing strategy was higher than that of small AUC. And the mining results are obviously better than other module mining algorithm, module biological significance is significant. NMCOM algorithm can not only accurately and effectively mining prostate cancer candidate disease module. And can be extended to mining other disease candidate modules.
【作者单位】：西北工业大学自动化学院信息融合技术教育部重点实验室;
【基金】：国家自然科学基金资助项目(91430111,61473232,61170134) 西南财经大学“中央高校基本科研业务费专项资金-青年教师成长项目”(JBK150134)~~
【分类号】：R737.25
【正文快照】： 前列腺癌在许多西方国家是男性最常见的恶性肿瘤,占男性癌症死因的第二位[1],仅2012年美国就有超过250万的男性被诊断出患有前列腺癌[2].在我国,前列腺癌在男性泌尿、生殖系统恶性肿瘤中发病率已跃居第三位,成为一个不容忽视的问题[3].近年来,前列腺癌的医疗诊断和治疗技术有

【参考文献】