
Brain Morphometry BIRN — Workflows
In managing analyses and processing of data, there are two important activities related to workflows: first, the management of encapsulated workflows within a common framework and second, the creation and running of specific workflows and application pipelines.
Workflows
The LONI Pipeline uses a simple, graphical environment to allow the user to easily buildup a complex data processing stream. The interface is intuitive and allows the user to build up their processing stream from a diverse set of programs. In addition, the Pipeline will automatically parallelize programs whenever possible to maximize efficiency when supercomputing resources are available. The LONI Pipeline has library definitions available for: AIR, MNI, BrainSuite, FSL, FreeSurfer, ITK, VTK, 3D Slicer. The LONI Pipeline version 4.0 can be downloaded here.
Under Development:
Portal-based Workflow
The Morphometry BIRN will be providing a portal-based workflow environment in the coming months developed at the Massachusetts General Hospital (MGH). A description of this work has been published (Murphy, S.N., Mendis, M.E., Grethe, J.S., Gollub, R., Kennedy, D., Rosen, B.R.; A Web Portal that Enables Collaborative Use of Advanced Medical Image Processing and Informatics Tools through the Biomedical Informatics Research Network (BIRN), AMIA, Fall Symp. 2007: in press.). This system is based on a website that would accept uploaded images and host the calculating machinery so that imaging processing could occur on the website. Images would be returned once they had been processed, along with the numerical results of any calculations. The first test application developed by mBIRN for the Portal-based workflow is the gradient unwarping application (click here to learn more about what the mBIRN provides to correct for image distortions. The function of this application is to undo the image distortions present in brain structural images that arise from imperfections in the applied magnetic field gradients. The algorithm used in this process uses proprietary information from the scanner vendor with must be obtained through a research agreement. Intellectual property issues limit the usage of this application by the scientific community. Wrapping this application in a workflow should increase its usage by the neuroimaging community since it would hide the proprietary information from the user, thereby obviating the need for a research agreement.
The system relies on uploads and downloads to and from the BIRN Data Grid. The data grid provides a way to access data sets and resources based on their attributes and/or logical names rather than their names or physical locations and allows file security to be managed on a network shared resource in conjunction with the Grid Account Management Architecture (GAMA). The GAMA system is used for authorization and authentication and consists of two components, a backend security service that provides secure management of credentials and a front-end set of portlets and clients that provide tight integration into web/grid portals.

Fig. Schematic of IT infrastructure used with the portal based workflow. The workflow portlet processes data on the Data Grid using the Shared Resource Broker (SRB) and the processes are authenticated using the Grid Account Management Architecture (GAMA) server.
The main system software is divided amongst a Web Server and an Execution Server to comply with the general architecture of the BIRN portal. The Execution server has access to a (open source) Condor grid (http://www.cs.wisc.edu/condor/). The MGH group chose to use jBPM as the principle engine for scheduling and executing other applications because it is a reliable and open-source workflow engine that is particularly geared towards making human handoffs in a workflow. It’s “out of the box” functionality includes a set of services that allows breakpoints in a workflow to be defined such that the workflow will enter a “wait” state until human intervention occurs. This gives the chance for handoffs between groups to occur and intermediate calculations to be checked. Additional required software includes the web-portal open-source software GridSphere and the open source Apache Tomcat project.
Eclipse-based Workflow
The gradient unwarping workflow has also been implemented by the MGH group using java in the Eclipse environment. Eclipse is a free open-source java environment that allows the creation of modular java plug-ins to be put into a common framework. This development, parallel to that of the portal-based system, was undertaken to free the development of the workflow as much as possible from specific hardware and to allow the developers to work independently. Example screen shots are shown below. The use of java enables the developer to use plug-ins from existing repositories and tailor a front end to a specific need. It also has the advantage of being platform independent. These attributes also allow easy integration with HTML-based applications and the XNAT group is exploring ways of integrating this workflow environment with XNAT.

