Network Analysis Tools (NeAT) - Taverna workflows

NeAT Taverna workflows were developped with the help of Eric Vervisch.

The Taverna workbench provides tools to build, edit and browse workflows. Its graphical interface facilitates an easy building and running of workflows over distributed computer resources.

We present below some workflows. They use NeAT and other resources exposed as Web Services. They can be interconnected into reusable analysis workflows.

You can execute those workflows with your own data. Meanwhile data file examples are also available for each workflow.

HOW TO ? To execute these workflows in the Taverna Workbench, download the workflow (SCUFL XML format) on your machine and load it in Taverna with "File > Open workflow...". Alternatively, copy and paste the URL of the workflow download link into Taverna with "File > Open workflow location...". Sample input data are provided for each workflow. To use a sample input, download the file on your machine, then load this file in Taverna with "File > Run workflow >Load input".

How to find paths in protein-protein-interaction networks?

Yeast two-hybrid datasets are known to be very noisy. However, a protein-protein-interaction confirmed by several different experiments might be more reliable than one found in only one experiment. This example demonstrates a workflow that carries out path finding on the intersection network of two different yeast two-hybrid networks.

In this workflow, the following steps are performed: - first, the intersection between two undirected networks obtained from yeast two-hybrid data is computed with compare-graphs - second, Pathfinder is launched on the intersection network --- Download Workflow File (SCUFL) Download Sample Input (The sample input consists of the two yeast two-hybrid data sets from Uetz et al, 2001 and Ito et al, 2002, which are also available in the DEMO data section.)

How to identify biologically relevant clusters in biological networks?

This workflow forms a part of the Nature protocol presenting NeAT. In the first step, we extract clusters from a yeast-specific network obtained from STRING. Next, these clusters are compared to MIPS complexes, which comprise known physical interactions between proteins. We end up with a list that gives for each MIPS cluster the score of its overlap with clusters obtained from the STRING network.

In this workflow, the following steps are performed: - first, clusters are extracted from the input network using MCL. - second, the clusters are compared to known protein complexes obtained from MIPS with compare_classes --- Download Workflow File (SCUFL) Download Sample Input (The sample input consists of the STRING yeast network and the MIPS classes, both available in the Nature protocol data section.)

How to find paths in organism-specific metabolic networks?

Path finding in organism-specific metabolic networks allows to detect pathway variants that are specific to an organism or to a set of related organisms. KEGG PATHWAY is a database that offers organism-specific metabolic information.

In this workflow, the following steps are performed: - first, an organism-specific metabolic network is constructed using the KEGG network provider - second, paths are found in this network using Pathfinder --- Download Workflow File (SCUFL)

---