@MISC\{IMM2003-02419,
    author       = "M. E. Hansen and J. Smedsgaard",
    title        = "Automated data processing of high-resolution mass spectra",
    year         = "2003",
    month        = "apr",
    publisher    = "Informatics and Mathematical Modelling, Technical University of Denmark, {DTU}",
    address      = "Richard Petersens Plads, Building 321, {DK-}2800 Kgs. Lyngby",
    url          = "http://www2.compute.dtu.dk/pubdb/pubs/2419-full.html",
    abstract     = "There has been an almost explosive growth in performance and applications of Electrospray Ionization (ESI) Time of Flight (TOF) mass spectrometry, which today is one of the most efficient tools for screening of metabolites in complex bio-samples. Most efficiently {ESI-MS} can be used by directly infusion of crude extracts into the source taking advantage of the high sensitivity, high mass resolution and accuracy and the limited fragmentation. 

Unfortunately, there has not been a comparable development in the data processing techniques to fully exploit gain in high resolution and accuracy of the massive amounts of data. We present an automated data processing method to quantitatively compare large numbers of spectra from the analysis of complex mixtures, exploiting the full quality of high-resolution mass spectra. By projecting all detected ions - within defined intervals on both the time and mass axis – on to a fixed one-dimensional array, we obtain a vector that can be used directly as input in multivariate statistics or library search methods. 

We demonstrate that both cluster- and discriminant analysis as well as {PCA} (and related methods) can be applied directly on mass spectra from direct infusion analyses of crude extract to find the relationship between species from several species terverticillate Penicillium, and also that the ions responsible for the segregation can be identified. Furthermore the process can automate the process of detecting unique species and unique metabolites."
}