Structures Databases Personal

At present there are >100 protein kinase-inhibitor structures in publidy available structural databases which span 28 kinases and a variety of inhibitor structural classes. Based upon an analysis of these data a classification of ATP binding regions has been proposed by Traxler and Furet [59] and Bower (personal communication from Michael J. Bower, December 1999). In the subsequent discussion a slightly modified version of this classification will be used to organize the trends seen across kinases and inhibitor dasses (see Fig. 2.2). In the Traxler model, five sites were proposed of which three (adenine, sugar and phosphate-binding sites) can be directly related to ATP and two additional lipophilic sites which lay outside of the region occupied by ATP. In the Bower model, an additional polar site on the surface of the protein was proposed. 

Figure 1 shows a schema for an ideal distributed chemical information system. Several authors in this book refer to the need for standard interfaces. Ultimately, the personal computer will provide the graphics interface not only to personal computer databases but also to company databases running on the company mainframe, and possibly also through the same network to public hosts, so that the chemist using a personal workstation will be able to create queries which can be addressed to local files, company files and public files. Soon, chemical databases will be available on Compact Disk Read Only Memory (CD-ROM) searchable by both substructure and text. These too fit into the scheme of Figure 1. Databases such as infra-red spectra libraries will have structure-searchable components either on the personal computer or on the laboratory instrument and will also be used through the same graphical interface.

Table 9.6. Performance of CONCORD on a subset of the Cambridge Structural Database (X-ray structures) and the fine chemical directory (FCD) and modern drug data report (MDDR) . (Personal communication from R. S. Pearlman, College of Pharmacy, the University of Texas at Austin.)...

The World Wide Web has transformed the way in which we obtain and analyze published information on proteins. What only a few years ago would take days or weeks and require the use of expensive computer workstations can now be achieved in a few minutes or hours using personal computers, both PCs and Macintosh, connected to the internet. The Web contains hundreds of sites of Interest to molecular biologists, many of which are listed in Pedro s BioMolecular Research Tools (http // www.fmi.ch/biology/research tools.html). Many sites provide free access to databases that make it very easy to obtain information on structurally related proteins, the amino acid sequences of homologous proteins, relevant literature references, medical information and metabolic pathways. This development has opened up new opportunities for even non-specialists to view and manipulate a structure of interest or to carry out amino-acid sequence comparisons, and one can now rapidly obtain an overview of a particular area of molecular biology. We shall here describe some Web sites that are of interest from a structural point of view. Updated links to these sites can be found in the Introduction to Protein Structure Web site (http // WWW.ProteinStructure.com/). 

It is important to note that theoretic argument and empiric study have shown that the LOO cross-validation approach is preferred to the use of an external test set for small to moderate sized chemical databases [39]. The problems with holding out an external test set include (1) structural features of the held out chemicals are not included in the modeling process, resulting in a loss of information, (2) predictions are made only on a subset of the available compounds, whereas LOO predicts the activity value for all compounds, and (3) personal bias can easily be introduced in selection of the external test set. The reader is referred to Hawkins et al. [39] and Kraker et al. [40] in addition to Section 31.6 for further discussion of proper model validation techniques. 

Gower Publishing (Brookfield, VT) has published a series of electronic handbooks providing approved ingredient information on materials used in cosmetics, personal care additives, food additives, and pharmaceuticals. Academic Press, through its Sci-Vision branch, has just (2000) launched an ambitious service of CD-ROM-based toxicity database products which are structure and substructure searchable. 

Full-featured structure drawing and presentation program. Supports IUPAC nomenclature (names to structures, structures to names). Integrated with CBIS. Supports personal databases of structures, reactions, and graphics. 

---