VirtualPlant: a software platform to support systems biology research

Abstract

Data generation is no longer the limiting factor in advancing biological research. In addition, data integration, analysis, and interpretation have become key bottlenecks and challenges that biologists conducting genomic research face daily. To enable biologists to derive testable hypotheses from the increasing amount of genomic data, we have developed the VirtualPlant software platform. VirtualPlant enables scientists to visualize, integrate, and analyze genomic data from a systems biology perspective. VirtualPlant integrates genome-wide data concerning the known and predicted relationships among genes, proteins, and molecules, as well as genome-scale experimental measurements. VirtualPlant also provides visualization techniques that render multivariate information in visual formats that facilitate the extraction of biological concepts. Importantly, VirtualPlant helps biologists who are not trained in computer science to mine lists of genes, microarray experiments, and gene networks to address questions in plant biology, such as: What are the molecular mechanisms by which internal or external perturbations affect processes controlling growth and development? We illustrate the use of VirtualPlant with three case studies, ranging from querying a gene of interest to the identification of gene networks and regulatory hubs that control seed development. Whereas the VirtualPlant software was developed to mine Arabidopsis (Arabidopsis thaliana) genomic data, its data structures, algorithms, and visualization tools are designed in a species-independent way. VirtualPlant is freely available at www.virtualplant.org.

Figure 1.

Conceptual diagram of the VirtualPlant software system. VirtualPlant follows the e-commerce site logic. In e-commerce sites, users browse and query the database and add products of interest to their shopping cart. Users then check out and purchase the items in their cart. Similarly, VirtualPlant allows biologists to browse lists of genes or microarray experiments with desirable properties. Having found interesting data, they can load the data into the gene cart and “check out” to analyze the selected genes. Biologists can then analyze or visualize the data in the cart to generate biological hypothesis. Most tools in VirtualPlant can store their output in the Cart for a new round of analysis. This key feature allows for iterative filtering and refinement of large data sets.

Figure 2.

The VirtualPlant Web site. There are four main areas in the VirtualPlant Web site: (1) the navigation window (top), (2) the cart window (left), (3) the database browser window (bottom left), and (4) the analysis window (center). The navigation window contains links to the different contents in VirtualPlant. The cart window displays the contents of the cart, which are lists of genes and experiments that have been created and saved by the user. The database browser window allows the user to navigate through different types of data stored in the database. Clicking on “analyze” in the navigation window loads a detailed view of the cart in the analysis window where the user can select the gene or experiment and the different visualization and analysis tools from the pull-down menu.

Figure 3.

Genes correlated to NIA1 and their gene ontology annotations. A, Histogram representing the number of probes correlated with the NIA1 Affymetrix probe (259681_at). Orange bars represent the number of probes in the different correlation cutoff intervals. These can be selected by the “range to graph” sliding tool on the right of the graph. Clicking on the bars will display probes from the selected interval in the table below. B, Pie-chart of the gene ontology terms associated to 23 genes selected in A. Each term has a different color in the pie chart. The legend to the right of the pie chart indicates the name of the GO term. The pie chart is generated by selecting the genes from the table and clicking on the “pie” button at the bottom.

Figure 4.

BioMaps results of genes that are induced by nitrate in both the wild type and NR-null mutant. BioMaps graphical output is a directed acyclic graph that shows the functional terms that are overrepresented in the gene list analyzed. The gray nodes contain the genes annotated to a functional term. The other colored nodes of the graph correspond to functional terms. The colors indicate the statistical significance of the overrepresentation as indicated in the legend included in the figure. For example, orange nodes correspond to functional terms overrepresented with P ≤ 1e-10.

Figure 5.

Super node and gene network forms. Super node analysis groups the genes based on the biological processes, functional terms, and annotations associated with the genes. A, The super node network form allows the user to choose from a selection of different functional term annotations and the depth of the annotation. In this case, the grouping is based on “KEGG pathway and gene families,” and only the “direct associated” annotations are used. In the super node analysis, interactions between the biological processes are determined by the multinetwork data. Therefore, super node analysis will prompt the user with two forms: the super node network form and the multinetwork form. B, The gene network form allows the user to select from the different molecular interactions that are present in the multinetwork (see Table I for the list of resources available). In addition to the super node analysis, this form is also used for the network statistics tool.

Figure 6.

Super node network analysis of genes differentially expressed during seed development. The super node network graph allows the user to visualize relationships between biological processes. The nodes in the graph correspond to the super nodes, each grouping genes with common features, and edges connecting the nodes represent the different interactions between the genes in the super nodes (see text for details). Edge colors represent different interactions: blue edges, protein-protein interactions; black arrows, metabolic reactions; red arrows, predictions for transcriptional induction; and green arrows, predictions for transcriptional repression. The network shows “nitrogen metabolism” and its first neighbors in the super node seed-regulated network. The neighbors are mostly transcription factor families and two metabolic processes. The number near each name identifies the number of genes in the super node.

Figure 7.

Gene network analysis of genes differentially expressed during seed development. The gene network graph shows interactions between genes, gene products, and/or metabolites. Orange circles represent metabolites, green triangles represent transcription factors, purple diamonds represent microRNAs, and blue squares represent metabolic genes. Edge colors represent different interactions: blue edges, protein-protein interactions; black arrows, metabolic reactions; red arrows, predictions for transcriptional induction; and green arrows, predictions for transcriptional repression. Different miR164 genes are shown targeting two transcription factors that are indirectly connected to the metabolic genes. Out of the seven nitrogen metabolic genes present in this network, only ASN1 and ASN2 have predicted regulators based on correlated transcription analysis and predicted cis-element binding sites.