Abstract
Over the past few decades there has been an increased interest in using various analytical techniques for detecting and identifying microorganisms. More recently there has been an explosion in the application of matrix assisted laser desorption ionization time of flight mass spectrometry (MALDI-TOF-MS) for bacterial characterization, and here we optimize this approach in order to generate reproducible MS data from bacteria belonging to the genera Bacillus and Brevibacillus. Unfortunately MALDI-TOF-MS generates large amounts of data and is prone to instrumental drift. To overcome these challenges we have developed a preprocessing pipeline that includes baseline correction, peak alignment followed by peak picking that in combination significantly reduces the dimensionality of the MS spectra and corrects for instrument drift. Following this two different prediction models were used which are based on support vector machines and these generated satisfactory prediction accuracies of approximately 90%.