With the release of VisANT 4.06, users can easily perform a disease enrichment analysis on any input gene sets to find out whether specific diseases are statistically over-represented in a given set of genes (represented by a metanode in VisANT). In the following tutorial, we will use the breast cancer driver genes published by Gray, Joe and Brian[1] as an example to demonstrate how to determine the overrepresentative diseases in VisANT.


Step 1. First, load gene set/network into VisANT. Here, total 40 driver genes are imported and formed a signal metanode in VisANT. Multiple gene sets can be imported into VisANT at the same time. In this case, each gene set must be presented as an individual metanode in VisANT.



Step 2. If the imported genes do not contain any disease information, then we will need annotate those genes using VisANT platform. First, select all genes and then select [ Node(s) ] -> [ Disease Annotation ] -> [ Using Most Specific Disease Terms ] from the right click menu to perform the disease annotation function for all genes. VisANT will automatically annotate genes by using the disease-gene associations available in the public databases such as KEGG, GAD, OMIM and PharmGKB.




Step 3. The default FDR cut-off value of the enrichment analysis is 0.01. Before starting the analysis, we can adjust this value from the property-sheet




Step 4. The disease enrichment analysis can be started by clicking on [ MetaGraph ] -> [ Predict Associated Disease of Metanodes ] -> [ Using Hypergeometric Test over All Diseases databases] from the menu. When there are more than one metanode (gene set) available and none of them are selected by the user, VisANT will perform the enrichment analysis over all possible metanodes. The analysis of each metanode is independent to each other and the result will be listed separately in the final report. In contrast, if one or more metanodes are selected by the user, then VisANT will only perform the enrichment analysis on the selected metanode(s).



Step 5. The analysis will take a couple minutes depending on the size of the gene set. After finishing, the result will be shown in the browser. Only the over-represented diseases will be listed in the report and the diseases will be grouped by the database. In this example, not only breast cancer but also other types of cancers are enriched in this gene set. It is because many of genes in this genes set may also contribute to other cancers.



In this example, it is clear that the set of breast cancer genes may also involved in some other type of cancers, such as ovarian cancer, as shown in above figure.


Step 6. In some cases, a user may only want to test certain diseases. Those diseasjes need to be specified by selecting the disease(s) from the ‘Hierarchy Explorer’. Then select [ MetaGraph ] -> [ Predict Associated Disease of Metanodes ] -> [ Using Hypergeometric Test over selected Diseases databases] to perform the enrichment analysis. In this example, we will test if any GAD Cancer is enriched in our gene sets.



The enriched diseases will be listed in the browsers. Similar to the previous result, breast cancer and some of other cancers are also overrepresented.



[1] Gray, Joe, and Brian Druker. "Genomics: The breast cancer landscape." Nature 486.7403 (2012): 328-329.