Poster Sessions

BIOINFO-2

Viswanadham Sridhara

V. Sridhara, L. Geer, S. Bryant

Decreasing Database Search Times in ETD MS/MS Sequence Searching by Assignment of Parent Precursor Charge to MS/MS Spectra

Introduction: When the precursor charge is unmeasured, determination of the charge state promises to speed up database search times and, in cases, decrease false positives. An unmeasured precursor charge require database search algorithms to search for the entire possible precursor charge range, for example, 2+ to 7+ (or higher). In some versions of ETD MS/MS mass spectrometry, a series of charge reduced precursor peaks appears in the MS/MS spectra. Here, we show that discriminant analysis approach using the intensity and pattern of the series of charge reduced precursor peaks as input features allows us to assign a single correct charge or a smaller charge range to search, thus significantly decreasing database search times. Methods: The OMSSA algorithm is written in C++ using the NCBI C++ toolkit. Splus and Perl scripts are used for data analysis. ETD spectra were obtained using the Finnigan LTQ mass spectrometer (Thermo Electron, San Jose, CA). This spectrometer was equipped with a nano-flow HPLC microelectrospray ionization source and was modified to facilitate ETD. We thank the Hunt Lab at the University of Virginia for their assistance in gathering data for this study. Conclusion: Spectra obtained from ETD/ECD can contain charge reduced precursor peaks, usually of high intensity, and whose pattern is dependent on its parent precursor charge. We have analyzed the ion current and pattern of charge reduced series of precursor peaks from a set of MS/MS spectra, with normalized Total Ion Current (TIC). These spectra had known precursor charge and known true positive hits. Since the pattern of the charge reduced precursor peaks is unique for every parent precursor charge, the ion current in the bins associated with precursor peaks can identify the correct precursor charge or its possible range. An unusual feature of some ETD MS/MS spectra is that they contain series of precursor peaks which represent 2 different precursor charges. Here, We propose Linear Discriminant Analysis [LDA] approach to determine the charge state of the peptides using the intensity and pattern of charge reduced precursor peaks as input features. The conclusions of this study are outlined below: 1. Significant reduction in database search times compared to the more general range search method. 2. Increase in sensitivity is seen for <3%FDR rates. 3. For higher e-value hits (or for hits >3%FDR range), charge reduced precursor series peaks are not distinct or missing.