Bioinformatics Advance Access originally published online on May 3, 2005
Bioinformatics 2005 21(13):2957-2959; doi:10.1093/bioinformatics/bti467
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
DIG—a system for gene annotation and functional discovery
1Duke Institute for Genome Sciences and Policy, Duke University Medical Center Durham, NC 27710, USA
2Department of Molecular Genetics and Microbiology, Duke University Medical Center Durham, NC 27710, USA
3Institute of Statistics and Decision Sciences, Duke University Durham, NC, USA, 27708
*To whom correspondence should be addressed.
 |
Abstract |
|---|
 TOP Abstract REFERENCES |
|---|
Summary: We describe a database and information discovery system named DIG (Duke Integrated Genomics) designed to facilitate the process of gene annotation and the discovery of functional context. The DIG system collects and organizes gene annotation and functional information, and includes tools that support an understanding of genes in a functional context by providing a framework for integrating and visualizing gene expression, protein interaction and literature-based interaction networks.
Availability: The DIG system is freely accessible at http://dig.cagp.duke.edu/
Contact: mdelong{at}cgt.duhs.duke.edu
Supplementary information: http://data.cgt.duke.edu.dig
Genome-scale molecular profiling studies have been applied to an increasing number of applications addressing a wide variety of problems in biology, including subtle variation in tumors that define sub-types of significant clinical importance (Alizadeh et al., 2000; Bhattacharjee et al., 2001; Golub et al., 1999; Perou et al., 2000; Pomeroy et al., 2002; Ramaswamy et al., 2001; Shipp et al., 2002; Singh et al., 2002). Related studies with analysis methods aiming to predict phenotypes of unknown samples based on training data illustrate the great potential for genome-scale gene expression information in prognostic applications (Huang et al., 2003; Pittman et al., 2004; van de Vijver et al., 2002; van'T Veer et al., 2002; West et al., 2001). One key need is for integrated data exploration tools to access and explore biological information on genes playing roles in these analyses. To this end, we have developed an information resource—DIG (Duke Integrated Genomics)—that aims to assist in the discovery of functional context for genes identified in expression (and other) experiments. DIG provides tools that (1) integrate information about genes from different databases, (2) enrich Medline searches for relevant citations through custom queries, (3) organize datasets in shared workspaces and (4) link to graphical displays for visualization. The system is freely publicly accessible (https://dig.cagp.duke.edu/).
DIG links a user-defined list of genes to diverse databases to provide a comprehensive summary of known information about any given gene or pathway. The user can upload lists of genes to DIG through a web interface (Fig. 1A). DIG can accept GenBank accession numbers or gene or protein symbols. DIG then produces summaries of the sources of information via cross-references to external databases including genetic information (LocusLink, Homologene), gene function (Gene Ontology), involvement in pathways (BioCarta, KEGG, GenMAPP), relevance to disease (OMIM) and literature (Medline). Figure 1B gives an example for genes identified in a breast cancer ER (estrogen receptor) analysis.
|
DIG also provides tools for searching Medline to identify the co-occurrence of gene citations in published literature; this often reflects a biological relationship (Jenssen et al., 2001). Given a collection of genes identified in a profiling experiment, DIG first searches for all pairwise combinations of genes to identify relationships that may facilitate an understanding of the list (Fig. 1C). This can reveal processes that may not be obvious from independent queries for genes. DIG also performs pairwise searches of genes against user-defined terms to identify citations relevant to a specific aspect of gene function, such as involvement in a particular disease. The output is a list of titles along with links to the PubMed abstract. These searches make use of aliases of the gene names found in LocusLink. The search list, including the aliases, is generated automatically when the query is performed.
DIG provides additional access to three sources of networks to extend the analysis of a given gene through the generation of networks: large-scale gene expression datasets, protein interaction datasets and the published biological literature. Although the source of information in each case is distinct, the logic for identifying the relevant information is similar, as is, the mechanism for presenting the information. First, gene expression datasets provide the opportunity to identify functionally related genes based on statistical models of patterns of association in expression. We have developed such an approach using large-scale, sparse graphical models (Dobra et al., 2003; Jones et al., 2004 http://ftp.isds.duke.edu/WorkingPapers/04-06.html; Rich et al., 2005). This is a powerful statistical approach that has the potential to substantially advance our ability to identify functional relationships since it does not depend only on similarity of expression patterns; rather, this is a full statistical model that assesses the partial associations between one gene and another in the context of all others via sparse regression analysis, as examples in the references listed above illustrate.
Second, protein–protein interactions can represent enzyme–substrate interactions, regulatory interactions or the interaction of structural components within the cell. The ability to generate networks of these interactions (A + B and B + C equals A + C), rather than merely identifying the direct partners, has the potential to extend the understanding of functional networks beyond the immediate interacting proteins. The development of efficient, high-throughput methods for identifying protein interactions based on the yeast two-hybrid assay has facilitated the development of databases that describe these interactions.
Third, the published literature provides the opportunity to relate one gene to another—identifying functional relationships that will form the basis for establishing networks. The ability to identify publications that describe gene interactions or functional relationships between two or more gene products provides an opportunity to improve the understanding of expression profiling experimental results. DIG draws on a local copy of Medline to speed the processes of access, indexing, interrogation and searching. Since much of the basis of DIG is currently gene-centric, DIG has sub-indices containing instances of gene and protein names, and their variants over time, in the published literature. This information can represent gene and protein relationships within Medline in a manner similar to the representation of gene expression networks.
To manage and access the complexity inherent in the data that is rendered into networks, DIG has a core visualization facility (GraphExplore) that allows exploration of a visualized network with links to related online resources (Fig. 1D). GraphExplore is embedded as a core component for visualization within DIG, as well as acting as a stand-alone graphical visualization and exploration tool (Q. Wang, G. Yao, J.R. Nevins and M. West, submitted for publication, http://graphexplore.cgt.duke.edu). Users may create sub-network graphs by querying for lists of genes defined by statistically or biologically defined models. In addition to visualizing the network, GraphExplore provides a flexible, interactive annotation environment where each node can be linked out to its LocusLink description, and, in the case of gene expression networks, corresponding to Affymetrix probe set and chromosomal location information. GraphExplore can also integrate multiple networks, such as those generated under varying experimental conditions or those from different statistical models. Finally, GraphExplore produces a summary of the gene lists and statistical reports for a graph. It can export graphics in multiple output formats. GraphExplore is a full Java application and renders networks using Apache's Batik package (http://xml.apache.org/batik/); it can be launched from the main DIG query page and downloaded from the Duke website, together with full documentation and tutorial material.
The use of the various networks embedded in DIG to place a given gene into a functional context is illustrated in the Supplementary Material using an investigation of the Rb-E2F cell cycle pathway as an example.
Although other gene and pathway annotation tools have been developed in both academic and commercial settings, the DIG system provides an extremely useful combination of sources of gene annotation merged with the ability to explore complex networks that can extend the basic annotation analysis. The system has evolved out of project-based efforts to explore the significance of genes identified in gene expression profiling experiments and the need to have efficient and powerful mechanisms to place genes in a functional context. Such an integrative system facilitates the investigation of systems-level biological problems. The ability to understand these relationships, by identifying the relevant links that bridge distinct pathways, will be a critical next step in assimilating this knowledge into a unified, systems biology description of the activity within the cell.
Received on August 25, 2004; revised on April 23, 2005; accepted on April 23, 2005
|
REFERENCES |
|---|
|
TOPAbstractREFERENCES |
|---|
Alizadeh, A.A., et al. (2000) Distinct types of diffuse large B-cell lymphoma identified by gene expression profiling. Nature, 403, 503–511[CrossRef][Medline].
Bhattacharjee, A., et al. (2001) Classification of human lung carcinomas by mRNA expression profiling reveals distinct adenocarcinoma subclasses. Proc. Natl Acad. Sci. USA, 98, 13790–13795
Dobra, A., et al. (2003) Sparse graphical models for exploring gene expression data. J. Mult. Anal., 90, 196–212.
Golub, T.R., et al. (1999) Molecular classification of cancer: class discovery and class prediction by gene expression monitoring. Science, 286, 531–537
Huang, E., et al. (2003) Gene expression predictors of breast cancer outcomes. Lancet, 361, 1590–1596[CrossRef][Web of Science][Medline].
Jenssen, T.-K., et al. (2001) A literature network of human genes for high-throughput analysis of gene expression. Nat. Genet., 28, 21–28[CrossRef][Web of Science][Medline].
Jones, B., Carvalho, C., Dobra, A., Hans, C., Carter, C., West, M. (2004) Experiments in stochastic computation for high dimensional graphical models. Stat. Sci., in press.
Perou, C.M., et al. (2000) Molecular portraits of human breast tumors. Nature, 406, 747–752[CrossRef][Medline].
Pittman, J., et al. (2004) Models for individualized prediction of disease outcomes based on multiple gene expression patterns and clinical data. Proc. Natl Acad. Sci. USA, 101, 8431–8436
Pomeroy, S.L., et al. (2002) Prediction of central nervous system embryonal tumour outcome based on gene expression. Nature, 415, 436–442[CrossRef][Medline].
Ramaswamy, S., et al. (2001) Multiclass cancer diagnosis using tumor gene expression signatures. Proc. Natl Acad. Sci. USA, 98, 15149–15154
Rich, J., et al. (2005) Gene expression profiling and genetic markers in glioblastoma survival. Cancer Res., 65, 4051–4058
Shipp, M.A., et al. (2002) Diffuse large B-cell lymphoma outcome prediction by gene-expression profiling and supervised machine learning. Nat. Med., 8, 68–74[CrossRef][Web of Science][Medline].
Singh, D., et al. (2002) Gene expression correlates of clinical prostate cancer behavior. Cancer Cell, 203–209.
van de Vijver, M.J., et al. (2002) A gene-expression signature as a predictor of survival in breast cancer. N. Engl. J. Med., 347, 1999–2009
van'T Veer, L.J., et al. (2002) Gene expression profiling predicts clinical outcome of breast cancer. Nature, 415, 530–536[CrossRef][Medline].
West, M., et al. (2001) Predicting the clinical status of human breast cancer by using gene expression profiles. Proc. Natl Acad. Sci. USA, 98, 11462–11467
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||