Abstract. The analysis of complex signals as obtained by mass spectrometric measurements is complicated and needs an appropriate representation of the data. Thereby the kind of preprocessing, feature extraction as well as the used similarity measure are of particular importance. Focusing on biomarker analysis and taking the functional nature of the data into account this task is even more complicated. A new mass spectrometry tailored data preprocessing is shown, discussed and analyzed in a clinical proteom study compared to a standard setting. Keywords. similarity measures, functional data, proteomics, mass spectrometry, pre-processing, wavelet analysis, generalized peak list