ASTOR - Arthropod Structure revealed by ultrafast Tomography and Online Reconstruction

Development of an online portal for morphological studies

As part of the ASTOR project the SCC develops a cloud-computing based virtual analysis platform which is managed by users through an online portal. Users will be able to run their OpenGL-based analysis application within virtual machines.

The current situation is characterized by a limited number of analysis workstations. Users have to reserve the appropriate time and of course visit a localized computing room. Providing virtual infrastructure should significantly improve this condition. Typical users are scientists and students from various institutes involved in the field of the synchrotron community.

The scientific part of the project includes the integration of high-speed network technologies such as InfiniBand within the virtual machines as well as the evaluation of alternative storage options such as noSQL database technologies. Our goals here include the improvement of managing metadata, system performance and stability as well as procurement and operating costs.

Official Abstract:

X-rays and tomography provide the opportunity to visualize internal structures of optically dense materials in 3D. The invention of synchrotron-X-ray-microtomography was the onset of a new era of morphological research on microscopically small animals (e.g. micro-arthropods). Analyzing such 3D-data is time consuming and technically challenging. Especially the automation of classification processes needs close cooperation of biologists and image processing experts. Right now, a new high-speed-tomography setup (UFO) is built at the ANKA synchrotron in Karlsruhe (KIT). This setup will enable unrivaled opportunities of high-throughput measurements and 3D/4D-tomographic imaging of dynamic systems and living organisms. However, this setup results in such large amounts of data that technical limitations will be reached regarding data acquisition, storage, organization and analyses. Therefore, users will no longer be able to process the resulting data at their home institutions, even with highly equipped computers. Hence, a strong integration of knowledge from biology, image processing and data management is needed to enable this new and fascinating high-speed-option for regular users in the future. Using the most speciose animal group on earth (arthropods) as a model system, the network for functional morphology and systematics aims to establish and standardize measuring parameters for UFO to meet the needs of a broad range of biological research. This will be achieved by optimizing data acquisition and processing, semi-automation of data analysis (reconstruction and segmentation/classification) and the creation of an online-portal providing easy access, stereoscopic visualization and semi-automatic analyses of the data using cloud technologies.