Radviz可视化方法在基因表达数据分析中的应用

Radviz Visualization and Application on Gene Expression Data Analysis

目的介绍Radviz可视化的基本原理和方法,并将Radviz可视化应用于基因表达数据的分类和特征选择。方法以结肠癌基因表达数据为例,结合启发式搜索和Vizrank可视化评估,利用Radviz可视化实现基因表达数据的分类和差异基因排序。结果由Vizrank算法得到排序前100的Radviz可视化结果,最优的Vizrank得分为0.9491,并得到了17个用于可视化分类的差异基因,其中部分基因获得了生物学解释。结论 Radviz能够形象的呈现隐含在数据中的模式特征,较好地用于基因表达数据的可视化分类和差异基因筛选。

Objective Introduce the basic principle and method of Radviz visualization, and apply Radviz Visualization to classification and feature selection of gene expression data. Methods Using Radviz Visualization, heuristic search and Vizrank visualization assessment to ana- lyze previously published colon cancer microarray data. Results Vizrank obtained top 100 most differential Radviz visualization（ the highest Vizrank score 0. 9491 ） that include 17 biological significant genes. Conclusion Radviz visualization were able to intuitively obtain hidden patterns in data, it could be applied to classification and feature selection of gene expression data.