Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Pathway Miner: extracting gene association networks from molecular pathways for predicting the biological significance of gene expression microarray data

Pathway Miner: extracting gene association networks from molecular pathways for predicting the... Vol. 20 no. 13 2004, pages 2156–2158 BIOINFORMATICS APPLICATIONS NOTE doi:10.1093/bioinformatics/bth215 Pathway Miner: extracting gene association networks from molecular pathways for predicting the biological significance of gene expression microarray data Ritu Pandey, Raghavendra K. Guru and David W. Mount Bioinformatics Core, Arizona Cancer Center, University of Arizona, Tucson, AZ 85724, USA Received on December 17, 2003; revised on March 18, 2004; accepted on March 23, 2004 Advance Access publication May 14, 2004 ABSTRACT from three different open source pathway resources—KEGG Summary: We have developed a web-based system (Pathway at http://www.genome.ad.jp, BioCarta at http://www.biocarta. Miner) for visualizing gene expression profiles in the context com and GenMAPP at http://www.genmapp.org. Pathway of biological pathways. Pathway Miner catalogs genes based Miner is an interactive user-friendly application that allows on their role in metabolic, cellular and regulatory pathways. users to selectively view, analyze, interpret and download A Fisher exact test is provided as an option to rank path- pathway information and networks extracted from the under- ways. The genes are mapped onto pathways and gene product lying database for high-throughput analysis of gene expres- association networks are extracted for genes that co-occur in sion data. The relational database that stores the functional pathways. The networks can be filtered for analysis based on relationships resides on a 900 MHz processor Sun Fire 280R user-selected options. Unix Server that uses 700 GB RAID system for data storage. Availability : Pathway Miner is a freely available web access- The website is driven by the Apache server. MySQL is used ible tool at http://www.biorag.org/pathway.html as the relational database management system. Contact: mount@email.arizona.edu Pathway Miner is a web tool and uses a client applica- tion to access the pathway database, extract and display the The recent explosion in the number of gene expression experi- network. The client is written using the Java Swing API and ments and the need to extract useful biological information runs on platforms with Java run-time environment version 1.4 from the resulting volumes of data has presented a new chal- or higher. lenge, that is, to identify the most significant changes in The tool provides two options to analyze genes in the data- gene expression and to interpret the results in terms of biolo- set: (1) to search genes based on their associations in metabolic gical relationships. There are a number of web resources that and/or cellular and regulatory pathways from the pathways provide data annotation services for gene expression datasets resources and (2) in addition to the above, perform a statist- (Guffanti et al., 2002). Other tools mine gene ontology attrib- ical test and rank significant pathways based on their P -values utes [GoFish (Berriz et al., 2003), ChipInfo (Zhong et al., from each of the three resources. A one sided Fisher exact test is implemented in Pathway Miner. This test uses the script by 2003)] and pathways to provide a framework for visualizing Øyvind Langsrud at http://www.matforsk.no/ola/fisher.htm. the data (Dahlquist et al., 2002; Bouton and Pevsner, 2002) Both the options include a fold change cutoff filter that organization of expression profiles (Grosu et al., 2002) or is used for removing less interesting data points. Inter- interactive modeling (Toyoda and Konagaya, 2003) of meta- active HTML outputs of organized pathway profiles are bolic pathways. All of these resources have their own unique produced with options for studying the pathway maps and capabilities and are individually focused on specific functions exploring a graph-based gene association network that is or features. extracted from each of three internet pathway resources sep- We describe Pathway Miner, a novel tool that mines gene arately. The HTML outputs are arranged in multiple ways associations and networks in biological pathway information to (1) view sample-specific pathway profiles, (2) compare represented in the currently available resources. An integrated database at http://www.biorag.org contains metabolic, cellular profiles between samples, (3) reveal pathways that are the and regulatory pathways for human and mouse gene products most highly represented in the datasets or (4) list gene products that participate in single or multiple pathways. Expression values from as many as four samples can be To whom correspondence should be addressed. 2156 Bioinformatics 20(13) © Oxford University Press 2004; all rights reserved. Gene association networks from pathways Fig. 1. Examples from Pathway Miner. (a) Default layout of the graph network. (b) Network filtered for down-regulated genes. (c) Network filtered for nodes with chosen edge strength. Nodes are labeled using the gene names. compared on any pathway map from KEGG, BioCarta or and network browser are available at http://www.biorag. GenMAPP and their corresponding graph networks can be org/help.html#pathway and http://www.biorag.org/applets/ extracted. Starting with an organized profile of genes and network_browser_help.html pathways based on the user dataset, an association graph Gene association networks are very helpful for analyz- network for the genes whose proteins function together in ing pathway relationships among genes that are found to one or more pathway is produced. This graph network is be co-regulated in expression data, and for analyzing up- drawn as a two-dimensional layout using the Neato program and down-regulated pathways that have several participat- in the Graphviz software (www.research.att.com/sw/tools/ ing genes. Most importantly, the tool can assist in finding graphviz/) and is displayed in a network browser that runs as a all pathways that are perturbed by genes whose expression Java applet. Single network or multiple subnetworks of gene is varying under given conditions. Other tools that independ- associations are produced where the nodes in the graph repres- ently map expression data to individual pathway maps are ent genes. Nodes are colored based on the expression values available. However, Pathway Miner has the unique features of up- and down-regulated genes (Fig. 1a). The edges indic- of extracting information from multiple pathways in which ate relationships between the genes in a pathway. The edges gene associations are found and presenting a global picture of are given weights based on the number of pathways in which the behavior of genes and pathways. We expect that Pathway the associating nodes (gene products) co-occur together in the Miner will be a beneficial tool for biologists. selected resource (KEGG, BioCarta or GenMAPP). Networks can be produced for multiple experiments for comparing the ACKNOWLEDGEMENTS expression patterns of genes. These networks can be filtered based on various criteria such as displaying only the nodes The authors thank Nirav Merchant and Gavin Nelson of the that reflect up- or down-regulated genes (network of down- Bio Computing Facility at the University of Arizona for sys- regulated genes shown in Fig. 1b) or those that have edge tems support. We also thank Dr Sylvan Green for helpful strength of a certain value (Fig. 1c). The network browser consultation on statistical analysis. This research work was provides options for performing several other operations for supported by NCI grant P30CA023074 to the Bioinformatics analyzing the gene network. The details for Pathway Miner Core of the Arizona Cancer Center. 2157 R.Pandey et al. REFERENCES expression results into metabolic networks. Genome Res., 12, 1121–1126. Berriz,G.F., White,J.V., King,O.D. and Roth,F.P. (2003) GoFish Guffanti,A., Reid,J.F., Alcalay,M. and Simon,G. (2002) The finds genes with combinations of Gene Ontology attributes. meaning of it all: web-based resources for large-scale func- Bioinformatics, 19, 788–789. tional annotation and visualization of DNA microarray data. Bouton,C.M. and Pevsner,J. (2002) DRAGON view: information Trends Genet., 18, 589–592. visualization for annotated microarray data. Bioinformatics, 18, Toyoda,T. and Konagaya,A. (2003) KnowledgeEditor: a new tool for 323–324. interactive modeling and analyzing biological pathways based on Dahlquist,K.D., Salomonis,N., Vranizan,K., Lawlor,S.C. and microarray data. Bioinformatics, 19, 433–434. Conklin,B.R. (2002) GenMAPP, a new tool for viewing and Zhong,S., Li,C. and Wong,W.H. (2003) ChipInfo: software analyzing microarray data on biological pathways. Nat. Genet., for extracting gene annotation and gene ontology informa- 31, 19–20. tion for microarray analysis. Nucleic Acids Res., 31, Grosu,P., Townsend,J.P., Hartl,D.L. and Cavalieri,D. (2002) 3483–3486. Pathway processor: a tool for integrating whole-genome http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Bioinformatics Oxford University Press

Pathway Miner: extracting gene association networks from molecular pathways for predicting the biological significance of gene expression microarray data

Pandey, Ritu; Guru, Raghavendra K.; Mount, David W.
Bioinformatics , Volume 20 (13): 3 – May 14, 2004
Download PDF
3 pages

Loading next page...
 
/lp/oxford-university-press/pathway-miner-extracting-gene-association-networks-from-molecular-p0ti6X09Yg

References (7)

Publisher
Oxford University Press
Copyright
Bioinformatics 20(13) © Oxford University Press 2004; all rights reserved.
ISSN
1367-4803
eISSN
1460-2059
DOI
10.1093/bioinformatics/bth215
pmid
15145817
Publisher site
See Article on Publisher Site

Abstract

Vol. 20 no. 13 2004, pages 2156–2158 BIOINFORMATICS APPLICATIONS NOTE doi:10.1093/bioinformatics/bth215 Pathway Miner: extracting gene association networks from molecular pathways for predicting the biological significance of gene expression microarray data Ritu Pandey, Raghavendra K. Guru and David W. Mount Bioinformatics Core, Arizona Cancer Center, University of Arizona, Tucson, AZ 85724, USA Received on December 17, 2003; revised on March 18, 2004; accepted on March 23, 2004 Advance Access publication May 14, 2004 ABSTRACT from three different open source pathway resources—KEGG Summary: We have developed a web-based system (Pathway at http://www.genome.ad.jp, BioCarta at http://www.biocarta. Miner) for visualizing gene expression profiles in the context com and GenMAPP at http://www.genmapp.org. Pathway of biological pathways. Pathway Miner catalogs genes based Miner is an interactive user-friendly application that allows on their role in metabolic, cellular and regulatory pathways. users to selectively view, analyze, interpret and download A Fisher exact test is provided as an option to rank path- pathway information and networks extracted from the under- ways. The genes are mapped onto pathways and gene product lying database for high-throughput analysis of gene expres- association networks are extracted for genes that co-occur in sion data. The relational database that stores the functional pathways. The networks can be filtered for analysis based on relationships resides on a 900 MHz processor Sun Fire 280R user-selected options. Unix Server that uses 700 GB RAID system for data storage. Availability : Pathway Miner is a freely available web access- The website is driven by the Apache server. MySQL is used ible tool at http://www.biorag.org/pathway.html as the relational database management system. Contact: mount@email.arizona.edu Pathway Miner is a web tool and uses a client applica- tion to access the pathway database, extract and display the The recent explosion in the number of gene expression experi- network. The client is written using the Java Swing API and ments and the need to extract useful biological information runs on platforms with Java run-time environment version 1.4 from the resulting volumes of data has presented a new chal- or higher. lenge, that is, to identify the most significant changes in The tool provides two options to analyze genes in the data- gene expression and to interpret the results in terms of biolo- set: (1) to search genes based on their associations in metabolic gical relationships. There are a number of web resources that and/or cellular and regulatory pathways from the pathways provide data annotation services for gene expression datasets resources and (2) in addition to the above, perform a statist- (Guffanti et al., 2002). Other tools mine gene ontology attrib- ical test and rank significant pathways based on their P -values utes [GoFish (Berriz et al., 2003), ChipInfo (Zhong et al., from each of the three resources. A one sided Fisher exact test is implemented in Pathway Miner. This test uses the script by 2003)] and pathways to provide a framework for visualizing Øyvind Langsrud at http://www.matforsk.no/ola/fisher.htm. the data (Dahlquist et al., 2002; Bouton and Pevsner, 2002) Both the options include a fold change cutoff filter that organization of expression profiles (Grosu et al., 2002) or is used for removing less interesting data points. Inter- interactive modeling (Toyoda and Konagaya, 2003) of meta- active HTML outputs of organized pathway profiles are bolic pathways. All of these resources have their own unique produced with options for studying the pathway maps and capabilities and are individually focused on specific functions exploring a graph-based gene association network that is or features. extracted from each of three internet pathway resources sep- We describe Pathway Miner, a novel tool that mines gene arately. The HTML outputs are arranged in multiple ways associations and networks in biological pathway information to (1) view sample-specific pathway profiles, (2) compare represented in the currently available resources. An integrated database at http://www.biorag.org contains metabolic, cellular profiles between samples, (3) reveal pathways that are the and regulatory pathways for human and mouse gene products most highly represented in the datasets or (4) list gene products that participate in single or multiple pathways. Expression values from as many as four samples can be To whom correspondence should be addressed. 2156 Bioinformatics 20(13) © Oxford University Press 2004; all rights reserved. Gene association networks from pathways Fig. 1. Examples from Pathway Miner. (a) Default layout of the graph network. (b) Network filtered for down-regulated genes. (c) Network filtered for nodes with chosen edge strength. Nodes are labeled using the gene names. compared on any pathway map from KEGG, BioCarta or and network browser are available at http://www.biorag. GenMAPP and their corresponding graph networks can be org/help.html#pathway and http://www.biorag.org/applets/ extracted. Starting with an organized profile of genes and network_browser_help.html pathways based on the user dataset, an association graph Gene association networks are very helpful for analyz- network for the genes whose proteins function together in ing pathway relationships among genes that are found to one or more pathway is produced. This graph network is be co-regulated in expression data, and for analyzing up- drawn as a two-dimensional layout using the Neato program and down-regulated pathways that have several participat- in the Graphviz software (www.research.att.com/sw/tools/ ing genes. Most importantly, the tool can assist in finding graphviz/) and is displayed in a network browser that runs as a all pathways that are perturbed by genes whose expression Java applet. Single network or multiple subnetworks of gene is varying under given conditions. Other tools that independ- associations are produced where the nodes in the graph repres- ently map expression data to individual pathway maps are ent genes. Nodes are colored based on the expression values available. However, Pathway Miner has the unique features of up- and down-regulated genes (Fig. 1a). The edges indic- of extracting information from multiple pathways in which ate relationships between the genes in a pathway. The edges gene associations are found and presenting a global picture of are given weights based on the number of pathways in which the behavior of genes and pathways. We expect that Pathway the associating nodes (gene products) co-occur together in the Miner will be a beneficial tool for biologists. selected resource (KEGG, BioCarta or GenMAPP). Networks can be produced for multiple experiments for comparing the ACKNOWLEDGEMENTS expression patterns of genes. These networks can be filtered based on various criteria such as displaying only the nodes The authors thank Nirav Merchant and Gavin Nelson of the that reflect up- or down-regulated genes (network of down- Bio Computing Facility at the University of Arizona for sys- regulated genes shown in Fig. 1b) or those that have edge tems support. We also thank Dr Sylvan Green for helpful strength of a certain value (Fig. 1c). The network browser consultation on statistical analysis. This research work was provides options for performing several other operations for supported by NCI grant P30CA023074 to the Bioinformatics analyzing the gene network. The details for Pathway Miner Core of the Arizona Cancer Center. 2157 R.Pandey et al. REFERENCES expression results into metabolic networks. Genome Res., 12, 1121–1126. Berriz,G.F., White,J.V., King,O.D. and Roth,F.P. (2003) GoFish Guffanti,A., Reid,J.F., Alcalay,M. and Simon,G. (2002) The finds genes with combinations of Gene Ontology attributes. meaning of it all: web-based resources for large-scale func- Bioinformatics, 19, 788–789. tional annotation and visualization of DNA microarray data. Bouton,C.M. and Pevsner,J. (2002) DRAGON view: information Trends Genet., 18, 589–592. visualization for annotated microarray data. Bioinformatics, 18, Toyoda,T. and Konagaya,A. (2003) KnowledgeEditor: a new tool for 323–324. interactive modeling and analyzing biological pathways based on Dahlquist,K.D., Salomonis,N., Vranizan,K., Lawlor,S.C. and microarray data. Bioinformatics, 19, 433–434. Conklin,B.R. (2002) GenMAPP, a new tool for viewing and Zhong,S., Li,C. and Wong,W.H. (2003) ChipInfo: software analyzing microarray data on biological pathways. Nat. Genet., for extracting gene annotation and gene ontology informa- 31, 19–20. tion for microarray analysis. Nucleic Acids Res., 31, Grosu,P., Townsend,J.P., Hartl,D.L. and Cavalieri,D. (2002) 3483–3486. Pathway processor: a tool for integrating whole-genome

Journal

Bioinformatics – Oxford University Press

Published: May 14, 2004

There are no references for this article.