SEARCH OF SEQUENCE DATABASES WITH UNINTERPRETED HIGH-ENERGY COLLISION-INDUCED DISSOCIATION SPECTRA OF PEPTIDES

We have broadened the utility of the SEQUEST computer algorithm to per mit correlation of uninterpreted high-energy collision-induced dissoci ation spectra of peptides with all sequences in a database. SEQUEST no w allows for the additional fragment ion types observed under high-ene rgy conditions. We analyzed spectra from peptides isolated following t rypsin digestion of 13 proteins. SEQUEST ranked the correct sequence f irst for 90% (18/20) of the spectra in searches of the OWL database, w ithout constraint by enzyme cleavage specificity or species of origin. All false-positives were flagged by the scoring system. SEQUEST searc hes databases for sequences that correspond to the precursor ion mass +/- 0.5 u. Preliminary ranking of the top 500 candidates is done by ca lculation of fragment ion masses for each sequence, and comparison to the measured ion masses on the basis of ion series continuity, summed ion intensity, and immonium ion presence. Final ranking is done by con struction of model spectra for the 500 candidates and constructing/per forming of a cross-correlation analysis with the actual spectrum. Give n the need to relate mounting genome sequence information with corresp onding suites of proteins that comprise the cellular molecular machine ry, tandem mass spectrometry appears destined to play the leading role in accelerating protein identification on the large scale required. ( C) 1996 American Society for Mass Spectrometry