Automatic Microscopy in Biology

The Max Planck Institute of Molecular Cell Biology and Genetics (MPI-CBG) in Dresden conducts research on the molecular mechanisms of absorption and trans­por­ta­tion of molecules in cells. To examine the un­der­­lying cellular processes parts of cells are high­lighted. For such an experiment genomes are pre­pared on plates (usually many) in small wells (up to 384 per plate). Each of the wells contains a slightly different sub-experiment. The preparation of the plates can take up to several months. The reactions in the cells create light emissions which are observed with an automatic microscope. Several images (with a size of up to 10 MB) are taken at different depths and at ten different positions in each well (Figure 1).


Figure 1: Picture taken by automatic microscope showing highlighted parts of cells.

By now several tens of millions of images were acqui­red which total amount of more than 100 TB.
The images are analysed by the image analysis suite Motion Tracking, developed at MPI-CBG. High ranking publications, for example in the Nature magazine [BIO1], are the result. In the future, three-dimensional and time-resolved movies are planned, which will cause a vast increase in storage and analysis demands.

Image Analysis Workflow

The workflow is as follows [BIO2]. It is seamlessly integrated into the working environment of the biologists (Windows-based image analysis suite).

        • Images are acquired by the automatic microscope in large numbers.
        • The images are stored on a file server in distinct data sets.
        • The biologist selects files and loads them with the image analysis software. The templates for the job description files and run scripts are created by Motion Tracking including the paths of the input files.
        • The image for each job is synchronized to an iRODS server. Only new files are transferred. The files are stored in the same data set directory struc­ture in iRODS as the original data set on the file server.
        • A UNICORE job is submitted via the UNICORE command line client UCC.
        • A UNICORE/X manages and forwards the job to the cluster.
        • The job is scheduled on the destination cluster and the run script is executed.
        • The script pulls the input files from the iRODS server. It authenticates via a proxy certificate created by UNICORE in the job directory on the cluster.
        • The run script executes Motion Tracking on the HPC system to analyse the image.
        • After the analysis is finished the results are put back into the iRODS server.
        • Motion Tracking regularly checks for new results and fetches them via the iRODS irsync.
        • Motion Tracking checks the results for correctness.
        • Motion Tracking saves the results by integrating them into the original data set.




Figure 2: Image analysis workflow. 



[BIO1]       Collinet, C.; Stöter, M.; Bradshaw, C. R.; Samusik, N.; Rink, J. C.; Kenski, D.; Habermann, B.; Buchholz, F.; Henschel, R.; Mueller, M. S. & others, Systems survey of endocytosis by multiparametric image analysis, Nature, Nature Publishing Group, 2010, 464, 243-249


[BIO2]      Grunzke, R.; Müller-Pfefferkorn, R.; Markwardt, U. & Müller, M., Advancing Cutting-Edge Biological Research with a High-Throughput UNICORE Workflow, Schriften des Forschungszentrums Jülich IAS Series Volume 9, 2011, 35


TUD, ZIH: Richard Grunzke, Ralph Müller-Pfefferkorn

Copyright by SWM, KIT – Universität des Landes Baden-Württemberg und nationales Forschungszentrum in der Helmholtz-Gemeinschaft
Templates Joomla 1.7 by Wordpress themes free