An improved normalized signal to noise ratio method for irrelevant genes removing

With the development of microarray technology, a lot of gene expression datasets have been applied to cancer classification and biomarker detection. Most of these gene expression datasets have small number of samples and tens of thousands of genes, so irrelevant genes eliminating is an important stage of feature selection for microarray expression data analysis. In this paper, an improved global normalized signal to noise ratio (gn-SNR) method for irrelevant genes removing is proposed. The method eliminates irrelevant genes by considering mean value and standard deviation with global normalization of different samples. First, the contributions of mean value and standard deviation between two kinds of samples are measured. Second, the global normalized contribution of the two parts is calculated. Finally, a threshold is used to remove irrelevant genes for microarray expression data analysis. The proposed method is examined on microarray of Leukemia dataset, Prostate dataset and Colon dataset. The best accuracies of the method in these datasets are 96.07%, 91.56% and 94.50%, respectively. The experimental results show that the proposed method has a powerful capability of irrelevant genes eliminating for microarray expression data analysis.