Peptide Identification from Mass Spectrometry

Abstract

Mass spectrometry is nowadays the method of choice for protein characterization in proteomics. Computer algorithms and software have played an essential role in analyzing the large amount of mass spectrometry data produced in any proteomics experiment. The fundamental task of such analyses is to identify the peptide for each spectrum in the data. Such identification is called “database search” if it requires the assistance of a protein database, and called “de novo sequencing” if not. In the past 20 years, many database search software tools have been developed for peptide identification; and a particular one, Mascot, that was developed in 1999, became dominant in the market. While new tools were continuously published in the following decade, none has significantly improved Mascot. The situation was disrupted around 2010, when the field witnessed a flurry of new database search tools that significantly improved Mascot in terms of both accuracy and sensitivity. In the first part of the talk, the peptide identification problem will be introduced, and the history briefly reviewed. In the second part of the talk, some practical concerns for using the bioinformatics tools in a proteomics lab are discussed. Properly dealing with these concerns resulted into the significant improvement we witnessed in the past few years. The second part of the talk will be focused on the research conducted at the author’s own group.