Noise reduction in microarray gene expression data based on spectral analysis

In genetic research, microarray chip carries thousands of genome expression profiles which allow biologists to analyze some of the developmental processes of life, such as biological reactions due to specific influences and so on. A main challenge of DNA microarray analysis is to separate the main gene expression from experimental noise. In order to ensure the accuracy of the following analysis, an effective noise filtering scheme is needed. In this paper, we propose a strategy to remove noise from gene expression profiles based on an autoregressive model based power spectrum analysis combined with singular spectrum analysis. This method helps us to determine the power spectrum effectively such that we can easily reconstruct the noise filtered time series signal.