Wavelet change-point prediction of transmembrane proteins

Motivation: A non-parametric method, based on a wavelet data-dependent thre shold technique for change-point analysis, is applied to predict location a nd topology of helices in transmembrane proteins. A new propensity scale ge nerated from a transmembrane helix database is proposed. Results: We show that wavelet change-point performs well for smoothing hydr opathy and transmembrane profiles generated using different scales. We inve stigate which wavelet bases and threshold functions ave overall most approp riate to detect transmembrane segments. Prediction accuracy is based on the analysis of two data sets used as standard benchmarks for transmembrane pr ediction algorithms. The analysis of a rest set of 83 proteins results in a ccuracy per segment equal to 98.2%; the analysis of a 48 proteins blind-tes t set, i.e. containing proteins not used to generate the propensity scales, results in accuracy per segment equal to 97.4%. We believe that this metho d can also be applied to the detection of boundaries of other patterns such as G + C isochores and dot-plots.