A Model for the Use of Cloud Computing in a Structural Gene Annotation Workflow

In order to understand the function of a gene or set of genes, it must first be sequenced and annotated.   An important stage in this process is structural gene annotation, which involves two approaches: ab initio and extrinsic.  Separately, these approaches suffer from flaws, which are reduced when they are combined.  Gene annotation workflows provide a method of combining and automating analysis tasks involved in both approaches.  The amount of computing power required for this type of application is large, so grid computing is often utilized.  However, cloud computing offers solutions for many problems created by the Grid, including inefficient resource allocation and resource underutilization, excess overhead from ensuring software compliance, issues with dynamic scheduling and planning, and other problems created from the use of heterogeneous resources.  For these reasons, it is justifiable to create a model for a structural gene annotation workflow in a cloud environment that is simple and efficient.  The first step was determining the requirements for the model.  The next step was to explore all of the available solutions based on the requirements analysis and design the model.  The result is a structural gene annotation workflow built using the Taverna Workbench 2.x and deployed on a Nimbus science cloud.  Since the Taverna system can be deployed on a grid platform as well, future work will aim to implement this workflow in the science cloud and on a science grid and compare its performance on both platforms to determine if the cloud-based model is an efficient alternative.