PPINGUIN: Peptide Profiling Guided Identification of Proteins improves quantitation of iTRAQ ratios

BMC Bioinformatics

Methodology Article

Knut Reinert¹  Ralph Schlapbach²  Christian Panse²  Dorothea Rutishauser²  Alexandra Chadt³  Tanja Dreja³  Hadi Al-Hasani⁴  Frank Kleinjung⁵  Johannes Schuchhardt⁵  Chris Bauer⁵ 

[1] Department Computer Science and Mathematics, Free University of Berlin, Berlin, Germany;Functional Genomics Center, UNI ETH Zurich, Winterthurerstrasse 190, CH-8057, Zurich, Switzerland;German Institute of Human Nutrition, Department of Pharmacology, Arthur-Scheunert-Allee 114-116, 14558, Nuthetal, Germany;German Institute of Human Nutrition, Department of Pharmacology, Arthur-Scheunert-Allee 114-116, 14558, Nuthetal, Germany;German Diabetes-Center at the Heinrich-Heine-University, Düsseldorf, Germany;MicroDiscovery GmbH, Marienburger Str. 1, 10405, Berlin, Germany;

关键词: Protein Inference;    Unique Peptide;    Quantitative Proteomics;    Quantitation Ratio;    Peptide Profile;   

DOI  :  10.1186/1471-2105-13-34

 received in 2011-08-02, accepted in 2012-02-16,  发布年份 2012

来源: Springer

【 摘 要 】

BackgroundRecent development of novel technologies paved the way for quantitative proteomics. One of the most important among them is iTRAQ, employing isobaric tags for relative or absolute quantitation. Despite large progress in technology development, still many challenges remain for derivation and interpretation of quantitative results. One of these challenges is the consistent assignment of peptides to proteins.ResultsWe have developed Peptide Profiling Guided Identification of Proteins (PPINGUIN), a statistical analysis workflow for iTRAQ data addressing the problem of ambiguous peptide quantitations. Motivated by the assumption that peptides uniquely derived from the same protein are correlated, our method employs clustering as a very early step in data processing prior to protein inference. Our method increases experimental reproducibility and decreases variability of quantitations of peptides assigned to the same protein. Giving further support to our method, application to a type 2 diabetes dataset identifies a list of protein candidates that is in very good agreement with previously performed transcriptomics meta analysis. Making use of quantitative properties of signal patterns identified, PPINGUIN can reveal new isoform candidates.ConclusionsRegarding the increasing importance of quantitative proteomics we think that this method will be useful in practical applications like model fitting or functional enrichment analysis. We recommend to use this method if quantitation is a major objective of research.

【 授权许可】

CC BY   
© Bauer et al.; licensee BioMed Central Ltd. 2012