The main focus of Philipp Heusers research is on the development of methods for the interpretation of very low resolution macromolecular density data. Therefore sophisticated pattern recognition methods, e.g. based on 3rd order moment invariants or distance matrices are employed.
On the one hand these investigations aim at novel algorithms for the placing of known 3D structures into very low-resolution density maps (i.e. 6-20Å). The main advance of the new software will be, that it does require only a minimum of prior knowledge. It does not require any symmetry of the molecule under investigation, nor a segmentation of the map, nor a prior placement of the domains in the map.
On the other hand the possibilities to interpret such low resolution maps without knowing anything about the molcules in advance are exploited. Based on a number of features the low-resolution density is analysed for patterns of density variation that complies with known patterns from the PDB. Based on the congruence of patterns in the density maps and in the known data from the databases a model of the molecule is built, which in case of x-ray maps can also be used for the extension and improvement of the phases (cmp. Heuser et al. 2009).
Further research interests include the development of predictive methods for structural biology, e.g. for protein structure prediction and for the prediction of the quaternary structure of protein complexes (protein-protein docking).