Evaluation of the Consensus of Four Peptide Identification Algorithms for Tandem Mass Spectrometry Based Proteomics
- *Corresponding Author:
- Dr. James Lyons-Weiler, 2Bioinformatics Analysis Core
Genomics and Proteomics Core Laboratories
University of Pittsburgh, Pittsburgh, PA
E-mail: [email protected]
Received Date: November 04, 2009; Accepted Date: February 05, 2010; Published Date: February 05, 2010
Citation: Dagda RK, Sultana T, Lyons-Weiler J (2010) Evaluation of the Consensus of Four Peptide Identification Algorithms for Tandem Mass Spectrometry Based Proteomics. J Proteomics Bioinform 3: 039-047. doi: 10.4172/jpb.1000119
Copyright: © 2010 Dagda RK, et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
The availability of different scoring schemes and f ilter set- tings of protein database search algorithms has gre atly ex- panded the number of search methods for identifying candi- date peptides from MS/MS spectra. We have previousl y shown that consensus-based methods that combine thr ee search algorithms yield higher sensitivity and spec ificity com- pared to the use of a single search engine (individ ual method). We hypothesized that union of four search engines ( Sequest, Mascot, X!Tandem and Phenyx) can further enhance sen- sitivity and specificity. ROC plots were generated to mea- sure the sensitivity and specificity of 5460 consen sus meth- ods derived from the same dataset. We found that Ma scot outperformed individual methods for sensitivity and speci- ficity, while Phenyx performed the worst. The union con- sensus methods generally produced much higher sensi tivity, while the intersection consensus methods gave much higher specificity. The union methods from four search alg orithms modestly improved sensitivity, but not specificity, compared to union methods that used three search engines. Th is sug- gests that a strategy based on specific combination of search algorithms, instead of merely ‘as many search engin es as possible’, may be key strategy for success with pep tide iden- tification. Lastly, we provide strategies for optim izing sensi- tivity or specificity of peptide identification in MS/MS spec- tra for different user-specific conditions.