PI: An open-source software package for validation of the SEQUEST result and visualization of mass spectrum
© Qiao et al; licensee BioMed Central Ltd. 2011
Received: 9 March 2011
Accepted: 15 June 2011
Published: 15 June 2011
Tandem mass spectrometry (MS/MS) has emerged as the leading method for high- throughput protein identification in proteomics. Recent technological breakthroughs have dramatically increased the efficiency of MS/MS data generation. Meanwhile, sophisticated algorithms have been developed for identifying proteins from peptide MS/MS data by searching available protein sequence databases for the peptide that is most likely to have produced the observed spectrum. The popular SEQUEST algorithm relies on the cross-correlation between the experimental mass spectrum and the theoretical spectrum of a peptide. It utilizes a simplified fragmentation model that assigns a fixed and identical intensity for all major ions and fixed and lower intensity for their neutral losses. In this way, the common issues involved in predicting theoretical spectra are circumvented. In practice, however, an experimental spectrum is usually not similar to its SEQUEST -predicted theoretical one, and as a result, incorrect identifications are often generated.
Better understanding of peptide fragmentation is required to produce more accurate and sensitive peptide sequencing algorithms. Here, we designed the software PI of novel and exquisite algorithms that make a good use of intensity property of a spectrum.
We designed the software PI with the novel and effective algorithms which made a good use of intensity property of the spectrum. Experiments have shown that PI was able to validate and improve the results of SEQUEST to a more satisfactory degree.
With the booming scale of spectra data, various software have been developed to identify proteins, such as SEQUEST , Mascot , SONAR , TANDEM , OMSSA . Without a quantitative understanding of the spectrum generating process, the widely used database searching algorithms, such as SEQUEST and MASCOT, adopt a simple fragmentation model to predict the theoretical spectrum. For example, SEQUEST assumes that cleavage will occur at peptide bonds in a uniform manner and simply ignores the influence of neutral losses. This simple strategy tends to result in a significant deviation of the predicated spectrum from the experimental one.
In our previous studies, we designed a novel statistical model to determine some important factors that influence the global fragmentation  and proposed an EM method to derive the neutral loss (including ammonia loss and water loss) possibilities for amino acids . We have used this model to predict theoretical spectrum. Using the derived quantitative parameters, we could generate the intensities for primary peaks (peaks corresponding to ions b and y) by simulating the tendency of cleavage towards middle, and estimate the cleave preference for a specific peptide bond and the intensities for neutral loss peaks from derived probabilities for each amino acids. Experimental results have shown that this model could predict a more realistic spectrum. In addition, we have used this prediction model to distinguish the false positive peptide identification in SEQUEST's output. For each peptide sequence reported by SEQUEST, we used our model to predict the theoretical spectrum and validate the peptide identification results according to the similarity between the theoretical spectrum and the experimental counterpart. On both LTQ and QSTAR spectra sets, this technique has helped to distinguish the false positive identification of SEQUEST.
We integrated these algorithms into an open source package PI (Peptide Identifier), which can be freely downloaded from http://www.bioinfo.org.cn/MSMS/.
Fragmentation of an amino acid bond in peptide produced n-terminal and c-terminal ions. In our Model, we assume the intensity of the fragmentation is related to the type of fragmentation bond and the position of this bond in the peptide. We named the model fragmentation event model. In a peptide A1A2 ⋯ AL with L amino acids, we take P(Ai, Ai+1) as the effective factor for the amino acid bond type, and take fi as influential contribution when the bond lies in the i - th position in this peptide. So we can get an event v with a intensity α × P(Ai, Ai+1) × fi, and then, a event vector representing the fragment event of a peptide can be derived asV = v1, v2, ⋯, vL-1. Here, we solve a non-linear programming problem to train these parameters with a automatic built training set .
Meanwhile, PI also has an option to include EM algorithm method to gain the probabilities of dehydration and deamination . With these probabilities, we can explain one fragmentation to multiple types of ions besides the major b and y ions, e.g., and their isotopic ions. Therefore, we can get a spectrum with reasonable intensity for multiple types of ions.
We have chosen the Jensen-Shannon Divergence, , as the scoring function to evaluate the similarity between the experimental spectrum and the theoretical one which is used for the final validation.
PI was written in Java and it is system independent. PI takes the results of SEQUEST as input files, reads the output format (out) files and spectrum format (dta) files, and exports a pix file in XML format. The pix file includes the scores from SEQUEST and PI and protein information. When running the program, the user should choose the input files firstly, and then specify the training set scale and the filtration condition in the process task dialog box. After this procession, PI assigns high scores to both the credible matches and those correct matches which are difficult to evaluate. Using the analysis function, PI can directly display the result pix file in curves. Moreover, the pix file also can be easily used by other software for the XML structural format.
Results and Discussion
We used several data sets to evaluate our PI software program, e.g., Comp12vs12standSCX_LCQ, StrepPyogenes_FFE2_LTQ-FT, StrepPyogenes_OGE_LTQ, and Gygi's data , etc. The first three data sets are downloaded on PeptideAtlas , which include varieties of types of data from iontrap instruments.
Gygi's data set contains LTQ spectra data and QSTAR spectra which covers not only the spectra files, but also SEQUEST's results and Mascot's results from Gygi's lab for the convenience of an overall evaluation.
The performance of PI and SEQUEST on different data sets, i
Because of the intuitive and simple interface style, PI is easy and convenient to work with. A complete manual in portable document format (PDF or docx) is provided and is accessible via web pages on our web site.
We designed the software PI with the novel and effective algorithms which made a good use of intensity property of the spectrum. Experiments showed that PI could validate the results of SEQUEST and improve the results to a satisfactory degree.
Availability and requirements
Project name: PI
Project home page: http://www.bioinfo.org.cn/MSMS/
Operating system: platform independent
Programming language: Java
Any restrictions to use by non-academics: none
This work was supported by National Natural Science Foundation of China under grants 30800189 and Beijing Municipal Natural Science Foundation under grants 5102029 (All results are available from http://www.bioinfo.org.cn/MSMS/)
- Yates J III, Eng J, McCormack A: Mining Genomes: Correlating Tandem Mass Spectra of Modified and Unmodified Peptides to Sequences in Nucleotide Databases. Analytical Chemistry 1995, 67: 3202–3210. 10.1021/ac00114a016View ArticlePubMed
- Cottrell J, London U: Probability-based protein identification by searching sequence databases using mass spectrometry data. Electrophoresis 1999, 20: 3551–3567. 10.1002/(SICI)1522-2683(19991201)20:18<3551::AID-ELPS3551>3.0.CO;2-2View ArticlePubMed
- Field H, Fenyoe D, Beavis R: RADARS, a bioinformatics solution that automates proteome mass spectral analysis, optimises protein identification, and archives data in a relational database. PROTEOMICS 2002, 2: 36–208. 10.1002/1615-9861(200201)2:1<36::AID-PROT36>3.0.CO;2-WView ArticlePubMed
- Craig R, Beavis R: TANDEM: matching proteins with tandem mass spectra. Bioinformatics 2004, 20: 1466–1467. 10.1093/bioinformatics/bth092View ArticlePubMed
- Geer L, Markey S, Kowalak J, Wagner L, Xu M, Maynard D, Yang X, Shi W, Bryant S: Open Mass Spectrometry Search Algorithm. Journal of Proteome Research 2004, 3(5):958–964. 10.1021/pr0499491View ArticlePubMed
- Lin Y, Qiao Y, Sun S, Yu C, Dong G, Bu D: A Fragmentation Event Model for Peptide Identification by Mass Spectrometry. Research in Computational Molecular Biology, 12th Annual International Conference, RECOMB 2008, 4955: 154–166.View Article
- Sun S, Yu C, Qiao Y, Lin Y, Dong G, Liu C, Zhang J, Zhang Z, Cai J, Zhang H, Bu D: Deriving the Probabilities of Water Loss and Ammonia Loss for Amino Acids from Tandem Mass Spectra. J Proteome Res 2008, 7: 202–208. 10.1021/pr070479vView ArticlePubMed
- Elias J, Haas W, Faherty B, Gygi S: Comparative evaluation of mass spectrometry platforms used in large-scale proteomics investigations. Nature Methods 2005, 2: 667–675. 10.1038/nmeth785View ArticlePubMed
- Deutsch EW, Lam H, Aebersold R: PeptideAtlas: a resource for target selection for emerging targeted proteomics workflows. EMBO reports 2008, 9(5):429–434. 8 8 10.1038/embor.2008.56PubMed CentralView ArticlePubMed
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.