基于基因功能表达谱的疾病分类:抗基因缺失的稳健性

利用基因表达谱数据，按Gene Ontology基因功能分类体系，将基因模块化地组织到具有显著生物学意义的低维差异表达功能模块单元中，构造新的指标用于分类疾病样本，从而提出了基于功能表达谱的分析新途径。新算法可稳健地抗基因检测缺失，抗基因表达变异，抗检测误差，并可以显著地降低分类特征维数(参与疾病分类的基因数目)。采用淋巴瘤数据集，比较了基于功能表达谱和常规的基因表达谱的决策树分类器。结果显示，基于功能表达谱可以得到高准确度的疾病样本分类结果，能够直接从功能水平上给出相应的生物学解释。通过仿真分析，进一步显示了基于功能表达谱的分类方法具有抗基因检测缺失的稳健性。

Disease classification based on gene function expression profiles: robustness against gene absence

Function expression profiles (FEP), obtained by organizing the original gene expression profiles onto low-dimension differentially expressed functional modules using Gene Ontology, were proposed as new analysis indexes to classify disease samples in this paper. This new approach can be robustly against gene absence, gene expression variation and measurement errors, and can also significantly reduce dimensions. Using a lymphoma dataset, the classification performance of a decision tree based on FEP and the conventional gene expression profiles (GEP) was compared. The analysis results indicated that precise classification of disease types and biological function comprehension can be achieved directly with FEP. In addition, the property of robustness against gene absence of the FEP based classification method was further shown by simulation analysis.