论文部分内容阅读
目的 探讨在基因表达谱数据分析中用主成分分析方法结合层次聚类法与K -均值聚类方法两种分类方法对组织样品分类的分类效果。方法 用主成分分析方法对数据降维后进行聚类分析 ,与不经主成分分析直接聚类 ,并结合筛选与组织样品分型相关的基因的各种筛选水平 ,评价聚类效果。结果 用约当指数进行评价两种聚类方法 :经主成分分析后用提取的主成分聚类与不用主成分的直接聚类效果不同 ;不同筛选相关基因的筛选水平对聚类效果也有影响。结论 对组织样品做聚类分析时 ,主成分分析能提高聚类质量 ,合理地筛选差异表达基因的方法能提高聚类效果
OBJECTIVE: To investigate the classification of tissue samples by principal component analysis and hierarchical clustering and K-means clustering in gene expression profiling. Methods The principal component analysis (PCA) method was used to reduce the dimensions of the data after clustering analysis. The results were compared with those without principal component analysis. The screening results were screened by screening various genes related to the typing of tissue samples. The results were evaluated by using the index of approximation as two clustering methods. The principal component analysis and principal component analysis did not show significant difference between the two methods. The screening results of different screening related genes also had an impact on the clustering effect. Conclusion When performing cluster analysis on tissue samples, the principal component analysis can improve the quality of clustering, and the method of screening differentially expressed genes can improve the clustering effect