Fachbereich Veterinärmedizin


Service-Navigation

    Publikationsdatenbank

    MALDI-typing of infectious algae of the genus Prototheca using SOM portraits (2012)

    Art
    Zeitschriftenartikel / wissenschaftlicher Beitrag
    Autoren
    Wirth, Henry
    von Bergen, Martin
    Murugaiyan, Jayaseelan
    Rösler, Uwe
    Stokowy, Tomasz
    Binder, Hans
    Forschungsprojekt
    Pathogene Pflanzen: Epidemiologie und Virulenzmerkmale von Prototheken humaner und tierischer Herkunft
    Quelle
    Journal of Microbiological Methods; 88(1) — S. 83–97
    ISSN: 0167-7012
    Sprache
    Englisch
    Verweise
    DOI: 10.1016/j.mimet.2011.10.013
    Pubmed: 22062088
    Kontakt
    Institut für Tier- und Umwelthygiene

    Robert-von-Ostertag-Str. 7-13
    Gebäude 35
    14169 Berlin
    Tel.+49 30 8385 1845 Fax.+49 30 83845 1863
    email:tierhygiene@vetmed.fu-berlin.de

    Abstract / Zusammenfassung

    MALDI-typing has become a frequently used approach for the identification of microorganisms and recently also of invertebrates. Similarity-comparisons are usually based on single-spectral data. We apply self-organizing maps (SOM) to portray the MS-spectral data with individual resolution and to improve the typing of Prototheca algae by using meta-spectra representing prototypes of groups of similar-behaving single spectra.

    The MALDI-TOF peaklists of more than 300 algae extracts referring to five Prototheca species were transformed into colored mosaic images serving as molecular portraits of the individual samples. The portraits visualize the algae-specific distribution of high- and low-amplitude peaks in two dimensions. Species-specific pattern of MS intensities were readily discernable in terms of unique single spots of high amplitude MS-peaks which collect characteristic fingerprint spectra. The spot patterns allow the visual identification of groups of samples referring to different species, genotypes or isolates. The use of meta-peaks instead of single-peaks reduces the dimension of the data and leads to an increased discriminating power in downstream analysis.

    We expect that our SOM portray method improves MS-based classifications and feature selection in upcoming applications of MALDI-typing based species identifications especially of closely related species.