Abstract P326: An Innovative Peptide Spectral Library Search Engine for Cardiovascular Proteomics
We developed a peptide spectral library search engine for the cardiovascular community. Over 50,000,000 spectra obtained with LTQ-Orbitrap instrument on cardiac mitochondria and proteasome were analyzed, and 108,268 representative spectra were included in this organelle-based library. An improved dot product algorithm, slide dot product, was coded to query user spectra against spectra in the library. This procedure provides a solution to optimally balance the speed and sensitivity in peptide identification. In addition, an innovative noise decoy protocol was engineered to distinguish spectra correlation attributed to noise signals of the analytical instruments. The immediate advantages include an effective improvement of both specificity and accuracy in protein identification, overcoming a long-lasting concern in experimental spectra library, the possibility of propagating spectra with inaccuracy and error. With an independent test dataset collected by LTQ-Orbitrap (∼10,000 spectra), this novel search engine identified 20% more spectra matches within 10% analytical time compared to other existing theoretical database search route. Furthermore, the broad utility of this search engine was also demonstrated in characterizing spectra collected by LCQ and Q-TOF instrumentation. This peptide spectral library search engine has been incorporated into the Cardiac Organellar Peptide Atals Library (COPa Library) of NHLBI Proteomics Center at UCLA and will be made accessible at www.HeartProteome.org.
- © 2011 by American Heart Association, Inc.