Structure recognition, problems with

Adamson et al. studied the structural features of chemical compounds in a large computer-based file [9]. The features are based on substructural fragments of their chemical structures. Several applications with an automated extraction scheme of such a substructural descriptor has been reported in structure-property and structure-activity problems [10-13], Substructural descriptors have also been used for the comparison of structural similarity and the clustering of chemical compounds based on it [14-18], However, the analysis of structural features of the compounds is a process necessary for the recognition of similarity. 

What does this have to do with structure prediction In fact, we would like to solve a fold recognition problem (Fig. 7.3). For each domain, we consider several million possible sequences, and identify the most favorable. These provide a signature of the 3D domain structure, or fold. Indeed, if we consider a new protein sequence, for which the 3D structure is unknown, we can compare it to our database of computed sequences. If the new sequence is similar to one or more in our database, we can infer that it will adopt the same 3D stmcture. In effect, we have identified the fold of the new sequence, and this is the first step towards structure prediction by homology modeling (above) [7]. 

The small size of hevein (43 residues), and the ease of its availability by biochemical purification or methods of peptide synthesis make this domain an excellent model system for the study of carbohydrate recognition by proteins. Herein, and taking the hevein domain as a model, we focus on the study of those molecular-recognition features relevant for the interactions between carbohydrates and proteins. We detail all of the techniques that are instrumental for tackling this problem, and how these can strategically be combined in an efficient manner. Particular emphasis is placed on the acquisition and analysis of data at atomic resolution (by NMR and/or X-ray ), and how these structural data relate with thermodynamic and kinetic information in reaching an understanding of the forces and interactions that play decisive roles in the interactions between carbohydrates and proteins. 

These results cannot be explained with any of the older theories of olfaction whereas the Enzyme Model of Olfaction not only can do that effortlessly, but actueLLly allows to predict these effects on the basis of generally accepted principles of molecular biochemistry. The concept of "STRUCTURE RECOGNITION AS PERIPHERAL PROCESS IN ODOR QUALITY CODING" represents only the special application of a more general mechanism of structvure recognition in peripheral processes to the problems of quality coding in olfac-r tion. 

---