|
|
|
|
|
|
|
a structure. Further, even with great effort, some proteins are intractable. The computational methods of protein homology model building, secondary structure prediction, and de novo 3-dimensional protein structure prediction are often used to provide approximate models [2224]. |
|
|
|
|
|
|
|
|
Once the involved task of structure determination is completed, a protein can be studied using molecular graphics. The computational methods of force-field energy analysis and molecular dynamics simulations can then be used to learn about how the protein works and interacts with its substrate. Molecular docking can be used to computationally introduce new compounds into the protein in an effort to understand if they would function as drugs. De novo design methods are used to build, from scratch, tailor-made molecular models directly into the binding site [25]. |
|
|
|
|
|
|
|
|
Knowledge of protein structure is only the first step in understanding biochemistry, a multifaceted and dynamical series of events. Through molecular dynamics simulation molecular movement can be studied visually in a way unavailable by experiment [2628]. |
|
|
|
|
|
|
|
|
The design of a molecule that binds well to a protein, although a substantial milestone, is only the first step in designing a drug. Bioavailability, how well a drug travels through the body, is critical for drug efficacy. The solubility of compounds in the body's various compartments and its transport between these compartments are key components of bioavailability. A number of methods that attempt to predict solubility in water, partitioning between aqueous and membrane phases, and movement within phases have been proposed [29]. Toxicity and other negative side effects have also halted the development of many promising candidates. The challenge of toxicity prediction has been the focus of many laboratories. Often, these approaches are identical to the QSAR methods mentioned above, used for the prediction of biological activities. In many cases, side effects are, in fact, just other biological activities. Sometimes, however, a single toxic response is actually due to a number of biochemical or physical mechanisms, all of which must be considered in predicting a response. Because of this, toxicity prediction can be very challenging [30]. |
|
|
|
|
|
|
|
|
Sometimes drug discovery is assisted by studying compounds with similar molecular structures or biological activity. Often, when looking for a place to starta lead compounda program might want to examine and evaluate many possibilities to find an optimum combination of activity, side effects, and of course novelty, for patenting purposes. Millions of small drug-sized molecules (less than a few hundreds of atoms) are known. Most pharmaceutical companies have hundreds-of-thousands of proprietary molecules. Additionally, the rapidly emerging field of combinatorial chemistry is already yielding thousands of new compounds in previously unthinkably short times [3133]. Within these structures is a wealth of information and a company's proprietary molecules provide potentially unique and hence valuable information with which to com- |
|
|
|
|
|