Advances in many fields have led to an accelerated demand for computational tools which can be used to study proteins. Biophysical techniques such as X-ray crystallography and multidimensional NMR are heavily dependent on computers for data retrieval and analysis, as well as for model-building of protein structures from those data. These techniques have provided a wealth of information about protein structures, much of which can serve as a starting point for computational studies of the protein's dynamics, thermodynamics, and substrate interactions. Spectacular advances in gene cloning and sequencing have provided tremendous amounts of protein sequence data, which demand computational analysis. New sequences are often compared against a gigantic library of other sequences in hope of discovering any structural similarities and evolutionary relationships. Sequence data is accumulating even faster than structural data, so there is great demand for computational techniques which can provide structural informational about the protein using both sequence and homology data. The eventual goal is a technique for predicting the three-dimensional structure of a protein from the sequence alone. Advances in computer power and computational techniques and the promise of continual advances in the future, have encouraged scientists to believe that computational solutions to these and other problems are possible, today or in the near future.
A wide variety of computational techniques have been applied to the study of proteins. The area of protein structure prediction is particularly wide-open, because the task seems too daunting for more traditional techniques. Among the many theoretical approaches applied to this goal are neural networks, lattice simulations, structural profiles, and analysis of patterns in sequence alignments. The most popular techniques for studying proteins, however, remain molecular modeling and molecular mechanics. The two terms are often used interchangeably, but the former refers primarily to the graphical display of molecules and the manipulation of these structures to obtain structural insights, while the latter refers to underlying computational techniques for analyzing molecular structure, dynamics, thermodynamics, and other properties. Molecular modeling techniques are widely used for predicting protein structures from structural homology and understanding enzyme-substrate interactions. Molecular mechanics techniques, especially molecular dynamics, are used in a wide variety of applications. It is not now feasible to simulate the entire folding process of a protein using molecular dynamics (folding can take several seconds or even minutes, but molecular dynamics usually uses timesteps of 10 seconds), but the unfolding process can be studied to gain insight into intermediates in the folding process. Other applications of molecular dynamics to proteins include the calculation of relative free energies of substrate binding to enzymes and analysis of protein-solvent interactions.