FluTyper - An Algorithm for Automated Typing and Subtyping of the Influenza Virus from High Resolution Mass Spectral Data

The use of high resolution mass spectrometry has been shown to be a rapid and accurate method for the typing and subtyping of influenza viruses. The detection of peptides with unique theoretical masses enables unequivocal assignment of type and subtype for any given sample. This research further extends the use of mass spectrometry for typing and subtyping by the development of naïve Bayes classifiers for automated analysis of influenza proteolytic mass spectra.

At the bottom of this page you will find the FluTyper software. FluTyper has been developed and implemented to enable automated typing and subtyping of influenza by the analysis of proteolytic high-resolution mass spectra. The software accepts any mass spectrum in tab-delimited format (m/z in first column and intensity in second column).

An example of a tryptic digest of the nucleoprotein from an H3N2 influnza virus can be downloaded: Brisbane_H3N2_NP.asc.

The standlone executables for Windows and Linux along with the source code is available to non-commercial users upon request.

A full description of the algorithm and implementation is described in the following manuscript:

Wong J.W.H., Schwahn A.B. and Downard K.M. (2010) FluTyper - an algorithm for automated typing and subtyping of the influenza virus from high resolution mass spectral data. BMC Bioinformatics, 11:266

Parameters:

Peak cutoff (relative intensity) - Cutoff at which to filter low intensity peaks.

Po cutoff - Cutoff at which to ignore masses that have a low probability across all types or subtypes.

Peak matching tolerance (ppm) - Error tolerance for matching a theoretical mass to a peak in a mass spectrum.

Protein type - The expected protein analysed in the mass spectrum.

Missed cleavages - Maximum number of missed tryptic cleavages to anticipate.

Possible modifications - CAM-Cys, Met-Ox, Pyro-Glu, PAM-Cys.

Relaxed - if checked, the uniqueness of masses are not evaluated which can lead to increased false positive identifications. This option may be useful if high-resolution (i.e. < 5 ppm) cannot be achieve by the mass spectrometer being used.

Please note that FluTyper will take approximately 10-15 seconds to analysis a spectrum. No analyses should exceed a minute.