Fromdifferentialexpression电子书文献.pdfVIP

From differential expression analysis to differential coexpression and differential regulation analysis Dr. Yuan-Yuan Li 李园园 PI of Systems Biology Group Shanghai Center for Bioinformation Technology Expression data analysis strategies • 差异表达基因筛选 Differentially expressed genes, DCGs • 差异表达功能筛选 Differentially expressed functions • 聚类分析Clustering • 基因调控网络构建 Gene regulatory networks • 分类分析Classification 分类Classification • 定义：从训练数据中找出一组能够描述数据集合典 型特征的模型（或函数），以便能够分类识别未知 数据的归属或类别。 – {(x1, x2, x3, x4)/x13, x22, x32, x42} • 肿瘤分型诊断：根据已知肿瘤类型的样本的表达谱 数据来构建分类器，然后利用它对新的表达数据进 行分类分析，确定肿瘤的类型。 – 肿瘤分型：形态学细胞学分子生物学 • A pioneer work: Golub TR, et al., Science 1999 – the capability of a DE gene set for molecular classification of cancer Expression data analysis strategies • 差异表达基因筛选 Differentially expressed genes, DCGs • 差异表达功能筛选 coexpression Differentially expressed functions • 聚类分析Clustering • 基因调控网络构建 Gene regulatory networks • 分类分析Classification Coexpression analysis • Coexpression relationship (pairs) • Coexpression clusters (sets) • Coexpression networks (structured sets) – Single network analysis – Differential network analysis 相似性度量函数 • 欧式距离（Euclidean distance） • Pearson相关系数（correlation coefficient ） • 互信息（mutual information） • …… Coexpression relationships (pairs) direct correlation Transitive correlation : Zhou, X., M. C. Kao, et al. (2002). Transitive functional