Statistical Analysis of Gene Co-Expression Networks by Maximal Overlap Discrete Wavelet Transform

In this paper, the maximal overlap discrete wavelet transform is used for the analysis of gene co-express network. Combined cross-correlation and the multi-resolution of wavelets, a scale-specific correlation measure is proposed, which can capture the relationship of co-expression under time-delay and local time points. The scale-specific correlation measure provide a novel method to capture more biological knowledge. Based the scale-specific correlation measure all time series gene expressions are decomposed by maximum overlap discrete wavelet transform at scale 1-4. The gene co-expression networks at each scale are constructed. The statistical features include cluster coefficients, average shortest path length and degree distribution at each scale are studied. A scale-invariant or fractal behavior for yeast gene co-expression network is investigated.