We are interested in translating 3D arrays of real numbers (images), that are output of 3D reconstructions algorithms, into simpler structures that nevertheless capture the topological / geometrical essence of the objects in the images. These descriptors, called component trees, contains all the information on the relationships between connected components when the image is thresholded at various levels. Two of the potential component use is to be used as descriptors of volume in databases, such as Quaternary Protein Structure (QPS) and the Electron Microscopy Data Base (EMDB), and in the process of biomolecular fitting.
Combining high-resolution structural information from parts of a macromolecular assembly with low-resolution information from the whole macromolecular assembly has proven to be useful in understanding the function and evolution of macromolecular complexes. High computational cost, conformational changes and the insufficient features in the target density map are some of the challenges faced by docking methodologies. Our initial work on component trees of embedded digital pictures gave an indication that an interactive visual tool can be designed to understand biological structural information at different resolutions. The planned research aims to use a graph-algorithmic approach to investigate the relationship between component trees from density maps at different resolutions in order to compute a list of possible or even exact docking locations.
[To be updated]