Research in nanoscopy (e.g. electron microscopy) and imaging involves large datasets of images. Not only is there a data storage challenge, data management and processing requires high performance computing facilities and connectivity for access to data.
The main projects involves building a pipeline to construct 3D electron microscopy images in collaboaration with NECEN.
The e-BioGrid team is open to take dedicated Nanoscopy and imaging projects on board. Contact us
if you are involved in Nanoscopy or Imaging life science research and you
have a need for support in software of hardware infrastructure.
With powerful machines available at the Netherlands Centre for Electron Microscopy various forms of 3D reconstruction in EM can be realized. Reconstruction methods range from near-atomic resolution to the level of ultra-structure. These methods are computationally intensive. Parallelization and Grid computing have only been partially adopted in this field. Particle-reconstruction has been implemented on a cluster using the IMAGIC software. X-ray crystallography uses the CCP4-suite and this suite is partially made suitable for the Grid. Computerized EM tomography invokes computer clusters using the IMOD software. Here, special attention need be given to enhancement of the 3D reconstruction. The goal of this project is to bundle 3D-reconstruction tools into an e-science problem solving environment for nanoscopy that is Grid enabled.
applicant:
Fons Verbeek, Leiden Institute of Advanced Computer Science
results:
Currently, the commonly used open source EMAN and Imod software packages for single particle reconstruction and tomography respectively, as well as Imagic are available to run on the BiG Grid cloud. A web interface has been build for users to upload and analyse data. Further input from EM researchers for requirements is welcome. Storage capabilities and further IT infrastructure regarding accounts and data access are being reviewed in the perspective of the upcoming usage to the NeCEN microscope.
status:
ongoing
team:
Fons Verbeek, Floris Sicking, N. Pannu, Jan-Pieter Abrahams, Bram Koster
Advanced analyses, easier Electron microscopy is of invaluable importance for the study of the complex organization and the architecture of cellular structures. In recent years new and (mostly automated) electron microscopical techniques have been developed (like electron tomography (ET) and focused ion beam scanning electron microscopy (FIB-SEM) that provide us with 3-dimensional (3D) information of the cell. This added dimension gave us new insights in cellular structures, and the interrelatedness between organelles and cellular processes. Automated 3D analysis methods are still in their infancy. Data extraction is still mainly depending on manual segmentation techniques, and therefore time consuming and subjective.
Structured storage Modern 3D electron microscopic recording methods and techniques (not only (S)TEM Tomography, but also FIB-SEM, ILEM and SEM) will have to cope with ever growing amounts of data. This amount of data needs to be stored in a structured and well-organized manner in order to be and remain accessible to its users. Currently data-storage is done by the individual researcher in many different ways and forms. A better structured and more uniform way of storing data and organizing data-management is a precondition for the primary science case, but it also creates the opportunity for long term use and thus enabling future reuse of electron microscopy information by research institutes and their distant collaborators.
The primary objective of this BiG Grid project is, to improve computer intensive analysis methods - including 3D template matching of 3D electron microscopy tomography data. In addition, the improvements include providing broad access of these methods to the 3D electron microscopy community. The secondary objective is to implement a data storage system for 3D electron microscopy data.
In order to achieve these objectives, an intensive collaboration of expert partners is needed to establish an infrastructure for the analysis of 3D electron microscopy data.
This project contains 5 main activities 1. Reduce the total compute time and improve the compute capacity by implementation of existing 3D analysis algorithms on GPU's 2. Improvement of the 3D analysis process by implementation of additional analysis and information management tools, objectifying and improving reliability of 3D data mining of electron tomography volumes 3. Improvement of the accessibility by creating an intuitive user environment 4. Improvement of storage, retrieval, archival and the controlled sharing of 3D data for electron microscopy data-analysis 5. Ensuring continuity and availability of the created solutions during and after the duration of the project
This project stems from a previous collaboration (IOP: IGE03012/VL-E: CellTom) in a IOP genomics programme. The Sara e-science support team has assisted in creating the project proposal and has established the project organization.
The foundation of this project is a proposal by the 3D electron microscopy group of the Utrecht University. This proposal is supported by the Leiden University Medical Center.
