Gene Expression Profiling of Colorectal Tumors and Normal Mucosa by Microarrays Meta-Analysis Using Prediction Analysis of Microarray, Artificial Neural Network, Classification, and Regression Trees

<table>Stacked bar chart of the gene importance of 55 genes. The percentile ranks of scores (RS%) of 55 genes were derived from PAM, ANN, CART, and C5.0 methods. The higher RS% represents that the gene was more important in the classification of colorectal tumors and normal mucosa tissues. CV: coefficient of variance. The RS% of genes in different methods was calculated by four approaches noted with the suffix, Δ, importance, centroid, and #. Symbols represent the way to calculate RS%. Δ: the significant times of each gene in 1,000 times of bootstrapping. Importance: the node location of genes in decision trees. Centroids: centroid values of each gene in PAM. #: the relative importance values in ANN.</table>

Disease Markers

fig3

Figure 3

Figure 3: Gene Expression Profiling of Colorectal Tumors and Normal Mucosa by Microarrays Meta-Analysis Using Prediction Analysis of Microarray, Artificial Neural Network, Classification, and Regression Trees 

Figure 3 | Gene Expression Profiling of Colorectal Tumors and Normal Mucosa by Microarrays Meta-Analysis Using Prediction Analysis of Microarray, Artificial Neural Network, Classification, and Regression Trees 