Molecular modeling is a computational approach and the “instruments” used are computers, computer programs, and knowledge. Every modeling question requires a different computational approach as well as specific knowledge and understanding of the biological system under investigation. Therefore, direct contact with the experimentalists is essential. I use the Discovery Studio modeling package for some of the questions, e.g., structure analyses and comparative modeling. The MolFit and P&S-filter set of programs is used for docking. MolFit was originally developed at the Weizmann Institute of Science by M. Eisenstein, I. Shariv, and E. Katchalski-Katzir and is continuously being improved by M. Eisenstein and her students. The P&S-filter is an extension of MolFit, which dramatically improves our ability to identify correct docking models. The level of predicting success with MolFit and P&S-filter is demonstrated in the world-wide Critical Assessment of Protein Interactions (CAPRI) experiment, where I successfully compete with research groups around the world.
The ANCHORSmap package was recently developed by M. Eisenstein and her student. This set of program maps show preferred binding locations of single amino acids on protein surfaces. It helps detect putative binding sites and is useful in designing binding partners.
Another set of programs developed by M. Eisenstein and her students is the EM2EM and PDB2EM package, which allows fitting and docking of low-resolution electron microscopy maps to each other and to atomic resolution structures.
Very often several computational tools are combined in order to address a modeling question and new computer programs and scripts are written, specific to the case.
Computations are performed on two PCs in Miriam’s room, one of which is equipped with stereo-viewing glasses, and on the Chemfarm multi-processor computer.