Fuzzy C-means method with empirical mode decomposition for clustering microarray data

Microarray techniques have revolutionized genomic research by making it possible to monitor the expression of thousands of genes in parallel. Data clustering analysis has been extensively applied to extract information from gene expression profiles obtained with DNA microarrays. Existing clustering approaches, mainly developed in computer science, have been adapted to microarray data. Among these approaches, fuzzy C-means (FCM) method is an efficient one. However, microarray data contains noise and the noise would affect clustering results. Some clustering structure still can be found from random data without any biological significance. In this paper, we propose to combine the FCM method with the empirical mode decomposition (EMD) for clustering microarray data in order to reduce the effect of the noise. We call this method fuzzy C-means method with empirical mode decomposition (FCM-EMD). Using the FCM-EMD method on gene microarray data, we obtained better results than those using FCM only. The results suggest the clustering structures of denoised data are more reasonable and genes have tighter association with their clusters. Denoised gene data without any biological information contains no cluster structure. We find that we can avoid estimating the fuzzy parameter m in some degree by analyzing denoised microarray data. This makes clustering more efficient. Using the FCM-EMD method to analyze gene microarray data can save time and obtain more reasonable results.