Since the sequensing of the human genome, all human protein sequences are known. However only a small number of these have crystallised and had their 3D structure determined. However a structure of a homologous protein are available for most proteins. If the sequence homology between a template and a target protein is suficiently (usually above 20% but with few insertions and deletions) a reasonable structure of the target can be obtained by homology modeling.
Before the model is build one must obtain a sequence alignment of the protein of interest and a suitable template. This is done by writing out the one letter codes for each protein on separate lines so that identical or similar residues in the proteins are positioned in the same column.
Here is the alignment for the below homology model.
(Requires Java Virtuel Machine)
First step in the model building would be to mutate all residues in the template into the corresponding residues of the target protein. After this step the protein backbone and most of the sidechains would have exactly the same conformation as the template. Next step is to build insertions and properly join the residues around deletions. Exactly how this is done depends on the method of choise. Third step is usually sidechain optimisation followed by loop optimisation. Finally the whole protein structure is gently energy minimised.
This example display a homology model of the protein NK2 (Neurokinin 2). The model was build using a distance geometry algorithm similar to algorithms used for solving NMR structures. The template and the mutated temolate can also be displayed. Notice that the coordinates for the protein backbone is identical for the template and the mutated template.
Molecular Mechanics (MM)
Molecular Quantum Mechanics
Physical & Chemical properties