多因素癌症诊断分类模型的建立和应用  

Multi-factorial classification model in diagnostic research of cancer

作　　者：孙淳[1] 刘银坤[1] 

机构地区：[1]复旦大学附属中山医院肝癌研究所,上海200032

出　　处：《生命的化学》2010年第3期368-373,共6页Chemistry of Life

摘　　要：随着基因组学、蛋白质组学技术的运用,对癌症相关分子标志物的筛选,以及基于多个分子标志物的诊断分类模型,成为近年来研究癌症诊断问题的热门途径。本文介绍了如何从基因芯片和质谱等数据库中筛选有诊断作用的分子标志物,以及构建高敏感性、高特异性诊断分类模型的技术路线;并根据在各类肿瘤中的研究实例,分析各分类算法的特点。The main problem of cancer diagnosis is the lack of proper biomarkers,limiting the diagnostic accuracy.A novel idea to solve this problem is establishing multi-factorial classification models using a series of biomarkers discovered by genomic and proteomic approaches.Meanwhile,underlying these models,various mathematical algorithms are applied,in terms of data preprocessing,feature selection,classification and validation.This paper gives an overview of the basic concepts and some related studies in this novel area of multi-factorial classification models for cancer diagnosis.

关 键 词：分子标记物 生物信息学 分类模型 

分 类 号：R730.4[医药卫生—肿瘤]