Methodology

GO Explorer: A gene-ontology tool to aid in the interpretation of shotgun proteomics data

Paulo C Carvalho^1,2* , Juliana SG Fischer^2,3* , Emily I Chen^2,6 , Gilberto B Domont³ , Maria GC Carvalho⁴ , Wim M Degrave⁵ , John R Yates III² and Valmir C Barbosa¹

¹  Systems Engineering and Computer Science Program, Federal University of Rio de Janeiro, Brazil

²  Department of Chemical Physiology, The Scripps Research Institute, La Jolla, USA

³  Chemistry Institute, Federal University of Rio de Janeiro, and Rio de Janeiro Proteomics Network, Rio de Janeiro, Brazil

⁴  Carlos Chagas Filho Biophysics Institute, Federal University of Rio de Janeiro, Rio de Janeiro, Brazil

⁵  Oswaldo Cruz Institute, Laboratory for Functional Genomics and Bioinformatics, Rio de Janeiro, Brazil

⁶  Department of Pharmacological Sciences, Stony Brook University, Stony Brook, NY, USA

author email corresponding author email* Contributed equally

Proteome Science 2009, 7:6doi:10.1186/1477-5956-7-6

Published:	24 February 2009

Abstract

Background

Spectral counting is a shotgun proteomics approach comprising the identification and relative quantitation of thousands of proteins in complex mixtures. However, this strategy generates bewildering amounts of data whose biological interpretation is a challenge.

Results

Here we present a new algorithm, termed GO Explorer (GOEx), that leverages the gene ontology (GO) to aid in the interpretation of proteomic data. GOEx stands out because it combines data from protein fold changes with GO over-representation statistics to help draw conclusions. Moreover, it is tightly integrated within the PatternLab for Proteomics project and, thus, lies within a complete computational environment that provides parsers and pattern recognition tools designed for spectral counting. GOEx offers three independent methods to query data: an interactive directed acyclic graph, a specialist mode where key words can be searched, and an automatic search. Its usefulness is demonstrated by applying it to help interpret the effects of perillyl alcohol, a natural chemotherapeutic agent, on glioblastoma multiform cell lines (A172). We used a new multi-surfactant shotgun proteomic strategy and identified more than 2600 proteins; GOEx pinpointed key sets of differentially expressed proteins related to cell cycle, alcohol catabolism, the Ras pathway, apoptosis, and stress response, to name a few.

Conclusion

GOEx facilitates organism-specific studies by leveraging GO and providing a rich graphical user interface. It is a simple to use tool, specialized for biologists who wish to analyze spectral counting data from shotgun proteomics. GOEx is available at http://pcarvalho.com/patternlab webcite.