Fulltext

Clustering and Filtering Tandem Mass Spectra Acquired in Data-Independent Mode

Pak, Huisong ; Nikitin, Frederic ; Gluck, Florent ; Lisacek, Frederique ; Scherl, Alexander ; Muller, Markus

In: Journal of The American Society for Mass Spectrometry, 2013, vol. 24, no. 12, p. 1862-1871

Aggiungi alla tua lista
    Titre
    • Clustering and Filtering Tandem Mass Spectra Acquired in Data-Independent Mode
    Auteur
    • Pak, Huisong. University of Geneva, Geneva, Switzerland
    • Nikitin, Frederic. SIB Swiss Institute of Bioinformatics, University Medical Center, 1, Rue Michel-Servet, 1211, Geneva 4, Switzerland
    • Gluck, Florent. University of Geneva, Geneva, Switzerland
    • Lisacek, Frederique. SIB Swiss Institute of Bioinformatics, University Medical Center, 1, Rue Michel-Servet, 1211, Geneva 4, Switzerland
    • Scherl, Alexander. University of Geneva, Geneva, Switzerland
    • Muller, Markus. University of Geneva, Geneva, Switzerland
    Type de document
    • Postprint
    Langue
    • Inglese
    Publié dans
    • Journal of The American Society for Mass Spectrometry, 2013, vol. 24, no. 12, p. 1862-1871. Springer US; http://www.springer-ny.com
    Autre version électronique
    Identifiant OAI-PMH
    • oai:doc.rero.ch:316521
    Summary
    Data-independent mass spectrometry activates all ion species isolated within a given mass-to-charge window (m/z) regardless of their abundance. This acquisition strategy overcomes the traditional data-dependent ion selection boosting data reproducibility and sensitivity. However, several tandem mass (MS/MS) spectra of the same precursor ion are acquired during chromatographic elution resulting in large data redundancy. Also, the significant number of chimeric spectra and the absence of accurate precursor ion masses hamper peptide identification. Here, we describe an algorithm to preprocess data-independent MS/MS spectra by filtering out noise peaks and clustering the spectra according to both the chromatographic elution profiles and the spectral similarity. In addition, we developed an approach to estimate the m/z value of precursor ions from clustered MS/MS spectra in order to improve database search performance. Data acquired using a small 3 m/z units precursor mass window and multiple injections to cover a m/z range of 400-1400 was processed with our algorithm. It showed an improvement in the number of both peptide and protein identifications by 8% while reducing the number of submitted spectra by 18% and the number of peaks by 55%. We conclude that our clustering method is a valid approach for data analysis of these data-independent fragmentation spectra. The software including the source code is available for the scientific community. Figure ᅟ