文摘
Mass spectrometry is a well-known technology used for the analysis of pure compounds as well as mixtures, widely applied in large-scale studies such proteomic studies. The result of mass spectrometric analyses is a mass spectrum, a profile of mass/charge values and corresponding intensity values originated from the analyzed compounds. In the case of large-scale analyses, raw mass spectra comparisons are difficult due to different drawback typologies: data defects, unusual distributions, underlying disturbs and noise, bad data calibration. A bunch of data elaborations is essential, from data processing to feature extraction, in order to obtain a list of peaks from different mass spectra. In this work, a workflow has been developed to process raw mass spectra and compare the new tidy ones with the aim of defining a robust procedure, suitable for real applications and reusable for different kind of studies. A similarity measure has been used for comparison purposes, in order to verify similarity among replicates and differences among analyzed samples, and a clustering method has been performed on fish species, in order to discover how they cluster statistically. A case study is shown with the application of the processing method to data obtained from the analysis of different fish species.