Accuracy of Structure Prediction

The ability to predict structures can be important in the design of ligands. Chapter 7 deals with stereoselectivities predicted and interpreted by molecular mechanics. 

---

If we can clean up the structural database and extend the set of data on known structures, we can increase the accuracy of structure predictions. We can also do so by advancing the methods of statistical learning that underlie the prediction machinery. The progress in both areas will add up to improve the predictions. 

A number of other software packages are available to predict NMR spectra. The use of large NMR spectral databases is the most popular approach it utilizes assigned chemical structures. In an advanced approach, parameters such as solvent information can be used to refine the accuracy of the prediction. A typical application works with tables of experimental chemical shifts from experimental NMR spectra. Each shift value is assigned to a specific structural fragment. The query structure is dissected into fragments that are compared with the fragments in the database. For each coincidence, the experimental chemical shift from the database is used to compose the final set of chemical shifts for the... 

Computers have greatly enhanced the speed and accuracy of the predictive nature of QSAR, however they are only tools to aid investigation and elucidation of the relationship between structure and properties. A clear understanding of the system and the fundamental questions that need to be answered are still a requirement for future research. 

In addition to X-ray diffraction and NMR, which are direct techniques, methods based on the calculation of predicted three-dimensional structures of molecules in the range of 3 to 50 amino acids based on energy considerations are under rapid development. These approaches use what are commonly called molecular dynamics and energy minimization equations to specify the most probable conformation of polypeptides and small proteins. Often, when combined with information from other sources, such as X-ray crystallography or NMR studies, they have been demonstrated to be quite useful. However, when standing alone, their power and the accuracy of their predictive capability remains to be seen. 

We particularly welcome estimation methods that incorporate available experimental data, such as melting point. Engineers engage in a number of steps to quantify their confidence in an estimation method and to improve the accuracy of the prediction when working with new molecules. As shown in Table 2, a new molecule with the structure has a molecular weight of 163.606, a measured normal boiling point (T ) of 237.6 C, and a few vapor pressure data. Using an estimation method... 

Fundamental physicochemical properties (see Chapter 5) such as partition coefficient (log P) and distribution coefficient (log D) and pfCa are well predicted directly from chemical structure [42, 43]. Aqueous solubility may also be predicted reasonably well [44—46] (see Chapter 4), though there are warnings on the accuracy of solubility predictions for in-house pharmaceutical company compounds since these tend to be quite different chemical structures from those used to develop the commercial models [47]. 

A researcher determined the following amino acid sequence that is part of a water soluble protein GIGAELKVLKKGGPAEISWIKRKRQQ. Using Tables 19.4 19.5, predict the secondary structure of this stretch of the protein at STP when it is dissolved in water. How might you determine the accuracy of your prediction ... 

In the first case, the structural description of over 100 thienyl- and furyl-benzimidazoles and benzoxazoles was multivariately characterized to identify three latent variables. A set of 16 informative molecules was derived thereafter on applying a central composite design criterion in these latent variables to all the available structures. The data were analyzed by a linear PLS model, which permitted the optimization of three structural features out of four. The fourth one, the substituent linked to the homocyclic ring of the bicyclic system was finally optimized by the CARSO procedure in terms of the substituents PPs, predicting two new compounds as possible optimal structures. Indeed, later analysis revealed the accuracy of these predictions. 

---