海量基因特征中早期癌症生物标记识别仿真  

Early Cancer Biomarker Identification Simulation in Vast Amounts of Genetic Traits

作　　者：关伟[1] 陆慧娟[1] 

机构地区：[1]中国计量学院,浙江杭州310018

出　　处：《计算机仿真》2013年第10期403-405,442,共4页Computer Simulation

基　　金：国家自然科学基金(61272315);浙江省自然科学基金(Y1110342)

摘　　要：研究早期癌症生物标记准确识别的问题。早期的癌症特征中,癌症的生物表示特征由于变化很小,基因位特征识别筛选过程中,传统的模式识别方式在发现与癌症相关的重要基因集合中搜索,挖掘过程由于特征标记过小,造成可识别实验数据增加,识别能力大幅下降,识别的不准确。提出了一种人工免疫的早期癌症生物标记识别方法。建立向量空间模型,描述早期癌症生物标记的基因特征,利用人工免疫方法,对所有的抗体克隆副本进行统计,获取特征匹配度,利用匹配度与阀值的正比关系,优化早期癌症生物标记的识别过程。实验结果表明,算法进行海量基因特征中的早期癌症生物标记识别中,能够极大地提高识别的准确性。The accurate identification of early cancer biomarkers was researched. A recognition method for early cancer biomarkers with artificial immune was put forward. A vector space model was established to describe the ge- netic characteristics of early cancer biomarkers, artificial immune method was employed to count all antibody clone copies, and obtain the degree of features matching. The direct ratio of the matching degree to the threshold was uti- lized to optimize the identification process of early cancer biomarkers. The experimental results show that when the al- gorithm is used to identify the early cancer biomarkers of abundant genetic traits, it can greatly improve the accuracy of identification.

关 键 词：基因特征 早期癌症 生物标记 人工免疫 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]