DomainFinder is an interactive program for the determination and characterization of dynamical domains in proteins. A dynamical domain is a region in a protein which can move essentially like a rigid body relative to other regions. Many, but not all, proteins have dynamical domains, and if they do, the relative movements of the domains are usually related to the function of the protein. The identification of dynamical domains is therefore useful in understanding the function of the protein. However, there are other situations in which the knowledge of dynamical domains is helpful. In structure determination, it can help to predict whether complexation with a ligand, crystal packing, or other external influences can lead to important conformational changes. In protein engineering, it can indicate whether a given modification is likely to change the dynamical behavior of the protein. In experimental observations of protein motion, it can suggest regions of particular interest. In numerical simulations, it can point out the slow motions whose correct sampling must be verified.
DomainFinder is based on new theoretical methods which permit the identification of dynamical domains from a single conformation with almost negligible computational effort. Any state-of-the-art desktop computer should be sufficient for analyzing proteins of about 1000 amino acid residues in a few minutes, and even the largest known protein structures can be analyzed in less than an hour using common workstations or high-end PCs. DomainFinder can also determine dynamical domains from a comparison of two given conformations of the same protein.
DomainFinder is written in Python, a high-level object programming language that is particularly well suited to the demands of scientific computations. The speed-critical parts are implemented in C. For common operations it makes use of the Molecular Modeling Toolkit, a library of Python code for molecular modeling and simulation applications. The results of a domain analysis can be saved with all details in an MMTK data file, which permits all kinds of further analysis.