waveTM: Wavelet-based transmembrane segment prediction
waveTM is a web tool for the prediction of transmembrane segments in a-helical membrane proteins. A "sliding" window of 20 residues is used in order to calculate an average residue hydrophobicity profile, using a "hydrophobicity" scale. Discrete Wavelet Transform is applied on the average residue hydrophobicity signal and the different frequency coefficients produced are adaptively thresholded so that a denoised signal is reconstructed. A dynamic programming algorithm processes the denoised signal to provide the optimal model for the number, the length and the location of membrane-spanning segments. The end points of the predicted segments are extended to include flanking hydrophobic residues. Topology prediction can also be obtained in conjunction with OrienTM (Liakopoulos et al, 2001).
Analysis of a non-redundant test set, provides a ~95% per segment accuracy and ~90% per residue accuracy. These results are reported in detail in the respective web pages linked below.