A comparison study of wavelet transforms for the visualization of differentially methylated regions in DNA samples

DNA methylation analysis has become an important topic in the study of human health. DNA methylation analysis requires not only a specific treatment of DNA samples based on bisulfite, but also software tools for their analysis. Although many software tools have been developed and some tools for detecting differentially methylated regions (DMRs) in the DNA have been proposed, there is still a lack of tools for interactively displaying the DMRs found. In previous works, we proposed a new approach based on the Haar wavelet transform for the interactive visualization of DMRs at different scales. However, the nature of this wavelet produces signal overlapping which prevents an easy detection of visual differences between case and control signals in some cases. In this paper, we show a comparison study of different wavelet transforms which may solve this problem. The evaluation results show that some of the considered wavelet transforms are very prone to yield negative signal segments due to the signal inertia, which may lead to false DMR detection. However, the spline wavelet transform does not significantly suffer from this effect, clearly highlighting true DMRs. Also, the spline wavelet transform does not need to reconstruct the signal in the lower transformation levels, which improves interactivity.