The Contextual Hubs Analysis Tool is a simple tool to identify which hubs in a network are preferentially connected to nodes of interest, or "contextual" hubs.
Highly connected nodes are important to the topology of a network and have biological relevance. In a static network model of a biological process, a simple hub analysis can be misleading. The extent to which a node acts as a hub can change with biological context, and with increasing amounts of experimental data available there's value in determing which hub nodes behave as hubs in context.
CHAT was designed with gene expression analysis in mind, defining "contextual" nodes as those that are significantly up- or downregulated over the course of an experiment. The tool can be used for any context value, however, and the user can choose to upload their own network or provide a list of genes of interest (human, mouse, or bovine) for network analysis.
The tool will determine which nodes are contextual based on the context values provided, then find the hub nodes in the network. It then does an overrepresentation analysis for each hub, giving a p-value for each hub indicating whether it is connected to more contextual interactors than expected by chance, using the hypergeometric distribution. This calculation takes into account the number of contextual nodes in the network, the number of interactors (contextual and overall) the hub has in the network, and the number of possible interactors the hub could have. The last term is calculated using InnateDB's collected database of interaction data, which includes data from a variety of sources as well as its own curated interation database.
The Web client is an easy-to-use interface that allows the user to upload a network or supply a list of genes with context, and vary the thresholds for context and for how connected a node must be to be considered a hub. The fully-featured command-line script provides more flexible context range options as well as the ability to define context as the combination of two values (such as a p-value and a fold-change value).
The research leading to these results received funding from the European Union Seventh Framework Programme (FP7/2007-2013) PRIMES project grant FP7-HEALTH-2011-278568 & Science Foundation Ireland grant 09/IN.1/B2642.
Developed by the Lynn EMBL Australia Group.