Using the GeneWeaver Tools
To use the GeneWeaver Tools, you must first create a project with a list of gene sets to analyze. This can be done from the search results, or the "My GeneSets" page, listing GeneSets you've uploaded and those available to you through your groups. Use the check boxes to select GeneSets, and the "Add to Project..." box to either create a new project or add the sets to an existing set.
From the "Analyze" page, you can select projects (or individual GeneSets from project) using the checkboxes, and then click any of the button across the top to run the specified tools, which are described below.
The individual tool results pages sometimes have more options available, which can be changed from the basic defaults. You can access them by clicking "Tool options" at the top of the page.
- The HiSim Graph tool produces tree-like diagrams which describe the hierarchy of gene sets in the analysis. The top of the diagram represents more broad gene sets, getting more specific as you move down the tree until arriving back at the original data sets. Nodes are formed by finding the common genes at each size of intersection of gene sets.
- The GeneSet Graph tool displays the bipartite graph of genes and gene set names.
- The Jaccard Similarity tool provides a pairwise comparison of the gene sets selected, showing the overlap of genes in the sets using venn diagrams. The nodes can also be clicked on to get the gene intersection lists (described below).
- The Jaccard Similarity coefficient (given in the basic table, and under each venn diagram) is calculated by taking the size of the intersection of the two gene sets, and dividing by the number of unique genes available in the two sets. A value of 1.0 means perfectly overlapping, and 0 means no intersection.
- The Hypergeometric Tests tool provides a pairwise comparison of the phenotypes selected, showing the overlap of genes in the sets using venn diagrams. The nodes can also be clicked on to get the gene intersection lists (described below).
- The Hypergeometric test produces 5 values: the hypergeometric score itself (hg), the odds ratio (or), upper tail p-value (ut), lower tail p-value (lt), and two-tail p-values (tt).
- The GeneSet Clustering tool performs clustering of the Jaccard similarity matrix across gene sets. Several different clustering algorithms are integrated into the system. You can select a method by expanding the tool options on the left sidebar, or on the results page. Use the pull down menu to choose a method.
- The clustering tool groups together gene sets based on Jaccard similarity, as a measure of gene set overlap. There are five different hierarchical methods used to cluster genes. The results of all five methods can be compared to measure the level of consensus between methods.
- The Triclique Viewer tool generates a graph using user-specified gene sets, and finds the maximal k-partite cliques within this graph to discover relationships between genes, gene sets, and ontology terms, and visualizes this in a chord diagram