Bioinformatics Advance Access originally published online on November 7, 2007
Bioinformatics 2008 24(8):1100-1101; doi:10.1093/bioinformatics/btm518
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Rintact: enabling computational analysis of molecular interaction data from the IntAct repository
1EBI-EMBL, Wellcome Trust Genome Campus, Hinxton, Cambridge CB10 1SD, UK and 2Computational Biology – FHCRC, 1100 Fairview Avenue North, M2-B876, Seattle, WA 98109, USA
*To whom correspondence should be addressed.
 |
ABSTRACT |
|---|
 TOP ABSTRACT 1 INTRODUCTION2 OBTAINING INTERACTION DATA3 COMPUTATIONAL ANALYSIS4 DISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
Motivation: The IntAct repository is one of the largest and most widely used databases for the curation and storage of molecular interaction data. These datasets need to be analyzed by computational methods. Software packages in the statistical environment R provide powerful tools for conducting such analyses.
Results: We introduce Rintact, a Bioconductor package that allows users to transform PSI-MI XML2.5 interaction data files from IntAct into R graph objects. On these, they can use methods from R and Bioconductor for a variety of tasks: determining cohesive subgraphs, computing summary statistics, fitting mathematical models to the data or rendering graphical layouts. Rintact provides a programmatic interface to the IntAct repository and allows the use of the analytic methods provided by R and Bioconductor.
Availability: Rintact is freely available at http://bioconductor.org
Contact: huber{at}ebi.ac.uk
|
1 INTRODUCTION |
|---|
|
TOPABSTRACT1 INTRODUCTION2 OBTAINING INTERACTION DATA3 COMPUTATIONAL ANALYSIS4 DISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
Protein–protein interaction mapping is a widely used approach to obtain a picture of cellular protein networks. The IntAct (Kerrien et al., 2006) database is a primary repository for the publication of molecular interaction data. There are many types of interactions, and each experiment is subject to effects that lead to error, so access to software tools for analysis and visualization is essential.
XML formats are intended for data exchange. They are usually not directly amenable for computational queries nor manipulations, and a transformation into data structures appropriate for the analysis of interest is needed.
We describe the Bionconductor package Rintact, which provides a programmatic interface to IntAct. It translates the primary data encoded in PSI-MI XML2.5 (Kerrien et al., 2007) files into R graph objects (R Development Core Team, 2007), which can then be analyzed by a variety of computational methods (Barenco et al., 2006; Chiang et al., 2007; Gentleman et al., 2004; Markowetz et al., 2005; Radivoyevitch, 2004; Siek et al., 2000–2001).
|
2 OBTAINING INTERACTION DATA |
|---|
|
TOPABSTRACT1 INTRODUCTION2 OBTAINING INTERACTION DATA3 COMPUTATIONAL ANALYSIS4 DISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
To illustrate the use of Rintact, we access the human CoIP data measured by Ewing et al. (2007) and the Y2H data by Stelzl et al. (2005). Files can either be downloaded and read from the local file system or read directly from the remote site; we construct the filename vectors for downloaded files:
> efiles = sprintf("human_ewing-2007-1_%02d.xml", 1:4) > sfiles = sprintf("human_stelzl-2005-1_1_%02d.xml", 1:2)
and convert the files into R intactGraph objects.
> ewingG = intactXML2Graph (efiles) > stelzlG = intactXML2Graph (sfiles)
Because both CoIP and Y2H use a bait/prey system, the resulting graph has directed edges from the bait to the prey.
To estimate the translation time of the function intactXML2Graph, we applied it to seven separate datasets from Utez et al. (2000) (two datasets), Cagney et al. (2001), Giot et al. (2003), Stelzl et al. (2005), Zhao et al. (2005) and Ewing et al. (2007). The data vary in size, and we found the general trend suggests a linear time algorithm based on the number of interactions. Thus Rintact provides a feasible approach in parsing the IntAct PSI-MI XML2.5 files.
IntAct uses internal, persistent identifiers called IntAct accession codes to unify the various identifier schemes of submitted datasets. The PSI-MI XML2.5 files provide translation information from the contained IntAct accession codes to various other commonly used molecule identifiers. This information allows the rendering of the interaction datasets using different types of molecule identifiers.
> ID = nodes(ewingG)[c(1, 45)] > translateIntactID(ewingG, ID, c("geneName", "uniprotId"))
geneName uniprotId EBI-1003700 "CENPH" "Q9H3R5" EBI-1046072 "PPP4C" "P60510"
The function intactXML2Graph can also be called on protein complex membership XML files, and the structure of the output is an intactHyperGraph. The relationship between proteins in multi-protein complexes is not binary; each protein complex can be represented as a hyperedge, and so the collection of protein complexes is a hypergraph.
|
3 COMPUTATIONAL ANALYSIS |
|---|
|
TOPABSTRACT1 INTRODUCTION2 OBTAINING INTERACTION DATA3 COMPUTATIONAL ANALYSIS4 DISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
After obtaining the molecular interaction data, we can exploit the various statistical methods in R and Bioconductor. For example, we can identify the densely connected subgraphs in Ewing et al.'s data using the highlyConnSG function from the RBGL package. Since highlyConnSG takes an undirected graph without self-loops, we first need to call the functions ugraph and removeSelfLoops on the directed data graph.
> g1 = removeSelfLoops(ugraph(ewingG)) > hc1 = highlyConnSG(g1)
A graph G with n vertices is highly connected if removal of any set of less than n/2 vertices does not disconnect G. Calling the length function on the first element of hc1 enumerates the number of highly connected subgraphs at 328, of which the largest has 640 vertices.
We can use the package ppiStats to compute summary statistics. Defining a viable prey (VP) as a protein that was found as a prey at least once in a given dataset (viable bait (VB) and viable bait/prey (VBP) are defined analogously (Chiang et al., 2007), we can produce the bar chart in Figure 1. It shows that Stelzl et al.'s (2005) Y2H data had a comparable number of viable baits to viable prey while in Ewing et al.'s (2007)'s CoIP experiment the viable prey population is larger than that of the viable baits.
|
We can view a subset of the CoIP data by rendering the subgraph induced by 10 baits and the group of preys they pull down in Figure 2 using Rgraphviz, and so we can easily see the clustering effects of the CoIP technology. Rintact can also work with the STRING database and the Cytoscape software via the Gaggle (Shannon et al., 2006) Bioconductor package. Other annotations can be obtained via the biomaRt (Durinck et al., 2005) Bioconductor package.
|
|
4 DISCUSSION |
|---|
|
TOPABSTRACT1 INTRODUCTION2 OBTAINING INTERACTION DATA3 COMPUTATIONAL ANALYSIS4 DISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
We have shown the capabilities of Rintact. While there are several software tools that also read PSI-MI XML2.5 files, Rintact has the additional benefit of being a computational conduit between IntAct and the analytic methods found in R and Bioconductor. Rintact provides an efficient and straightforward approach towards the analysis of molecular interaction data.
|
ACKNOWLEDGEMENTS |
|---|
|
TOPABSTRACT1 INTRODUCTION2 OBTAINING INTERACTION DATA3 COMPUTATIONAL ANALYSIS4 DISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
We would like to thank Abhishek Pratap and Li Wang for testing the Rintact software package. We acknowledge funding through the HFSP Grant RGP0022/2005 to W.H. and R.G. and NIH Research 1P41HG004059 to R.G.
Conflict of Interest: none declared.
|
FOOTNOTES |
|---|
Associate Editor: Alfonso Valencia
The authors wish it to be known that, in their opinion, the first two authors should be regarded as joint First Authors. 
Received on August 14, 2007; revised on October 10, 2007; accepted on October 10, 2007
|
REFERENCES |
|---|
|
TOPABSTRACT1 INTRODUCTION2 OBTAINING INTERACTION DATA3 COMPUTATIONAL ANALYSIS4 DISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
Barenco M, et al. Ranked prediction of p53 targets using hidden variable dynamic modeling. Genome Biol. (2006) 7.
Cagney G, et al. Two-hybrid analysis of the Saccharomyces cerevisiae 26s proteasome. Physiol. Genomics, (2001) 7:27–34.
Chiang T, et al. Coverage and error models of protein-protien interaction data by directed graph analysis. Genome Biol. (2007) 8.
Durinck S, et al. BioMart and Bioconductor: a powerful link between biological databases and microarray data analysis. Bioinformatics (2005) 21:3439–3440.
Ewing EM, et al. Large-scale mapping of protein-protein interactions by mass spectrometry. Mol. Syst. Biol. (2007) 3.
Gentleman RC, et al. Bioconductor: open software development for computational biology and bioinformatics. Genome Biol. (2004) 5:R80.[CrossRef][Medline]
Giot L, et al. A protein interaction map of Drosophila melanogaster. Science. (2003) 302:1727–1736.
Kerrien S, et al. IntAct – open source resource for molecular interaction data. Nucleic Acids Res. 35:D561–D565.
Kerrien S, et al. Broadening the horizon – level 2.5 of the HUPO-PSI format for molecular interactions. BMC Biol (2007).
Markowetz F, et al. Non-transcriptional pathway features reconstructed from secondary effects of RNA interference. Bioinformatics, (2005) 21:4026–4032.
Development Core Team R. R: A Language and Environment for Statistical Computing. (2007) Vienna, Austria: R Foundation for Statistical Computing. ISBN 3-900051-07-0.
Radivoyevitch T. A two-way interface between limited systems biology markup language and R. BMC Bioinformatics, (2004) 5:190–190.[CrossRef][Medline]
Shannon P, et al. The Gaggle: an open-source software system for integrating bioinformatics software and data sources. BMC Bioinformatics, (2006) 7.
Siek J, et al. The Boost Graph Library. (2000–2001) Cambridge: Cambridge University Press.
Stelzl U, et al. A human protein-protein interaction network: a resource for annotating the proteome. Cell, (2005) 122:957–968.[CrossRef][Web of Science][Medline]
Peter Uetz, et al. A comprehensive analysis of protein-protein interactions in Saccharomyces cerevisiae. Nature (2000) 403:623–627.[CrossRef][Medline]
Zhao R, et al. Navigating the chaperone network: an integrative map of physical and genetic interactions mediated by the Hsp90 chaperone. Cell, (2005) 120:715–727.[CrossRef][Web of Science][Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  B. Aranda, P. Achuthan, Y. Alam-Faruque, I. Armean, A. Bridge, C. Derow, M. Feuermann, A. T. Ghanbarian, S. Kerrien, J. Khadake, et al. The IntAct molecular interaction database in 2010 Nucleic Acids Res., January 1, 2010; 38(suppl_1): D525 - D531. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||