Max Planck Institute for Brain Research
While Cryo Electron Tomography offers an unique potential for direct visualization of the three dimensional structure of macromolecular compounds in their cellular environment, the severely noisy nature of the low dose images delivered by the microscope makes this imaging technique heavily dependent on a suitable computational treatment of the raw data. In particular, structure determination through Subtomogram Averaging requires the localization of large number of noisy copies of the compound of interest and extraction of the common signal through classification and alignment. This presentation focuses on several data processing challenges appearing in different stages of this technique, discussed in the context of the unified processing framework provided by the Dynamo software package. The new Catalogue module covers the logistical tasks attached to the organization of large datasets, providing specific tools for visualization and geometrical modeling of tomogram content by manual or semiautomated identification and annotation of structures as membranes, filaments or lattice-like arrangements of particles.
Further modules address the analysis of the identified subtomograms, guiding the users through the use of standardized analysis steps or rather supporting the design of new, case specific approaches. The software can harness High Performance Computing environments like CPU or GPU clusters, in order to cope with the numerically demanding nature of the technique, an aspect that will be showcased by different case studies.
Following this forum there will be a hands on demonstration of the Dynamo package, during which time attendees can see a preliminary introduction.