Pharmacophores representative queries

Finally, 3D pharmacophores can be used to provide a naturally partitioned space. By com bining the pharmacophore keys of a set of molecules one can determine how many of th potential 3- or 4- point pharmacophores are accessible to the set and easily identify thos which are not represented. This use of pharmacophores is the basis of a method namei Pharmacophore-Derived Queries (PDQ) [Pickett et al. 1996]. One feature of this particula method is that most molecules will occupy more than one cell (as nearly all molecules wil contain more than one 3-point pharmacophore due to the functionality present an( conformational flexibility). This contrasts with the usual situation, wherein each molecul occupies just one cell. 

The first implementations of 3D database searching software relied on the use of a single low-energy conformation to represent each molecule within a given database. - This was recognized as a serious deficiency, and new methodology was developed to obviate the problem. Attempts were made to construct pharmacophore search queries that implicitly took account of confor-... 

This method represents the most common and traditional application of computational tools to rational drug design. From a list of molecules of known activity, one can establish a 3D-pharmacophore hypothesis that is then transformed into a 3D-search query. This query is then used to search a 3D database for structures that fit the hypothesis within a certain tolerance. If the yield of active molecules is significant, then the query can be used to predict activities on novel compounds. In our situation, the enantiophore is built from the superposition of a list of sample molecules, which are all well separated on a given CSR Hence, the common features of this series of molecules can become a good enantiophore hypothesis for the enantiores-olution on this CSR... 

The SCF of the protein binding sites can be represented with colored bar-codes for which each color bit represents a 3D SCF pharmacophore feature. Binding sites with the highest number of matches between these triplets or quadruplets of SCF have the highest similarity with the query binding site (Fig. 5). The larger... 

The pharmacophore points in the Tripos implementation of DISCO, currently marketed under the name DISCOtech , can be represented as Tripos UNITY [56] query features and the models can be used directly for UNITY database searches or in combination with 3D QSAR such as CoMFA as described in [57]. 

The definition of pharmacophores is done manually by applying so-called schemes using a Pharmacophore Query Editor. A template molecule is generally used for this purpose. In the MOE environment, a scheme is a collection of functions that define how each ligand is annotated. This is accessed via an SVL function. The default scheme is called PCH (Polarity-Charged-Hydrophobicity). New schemes can be created to represent certain molecules better, e.g. Planar-Polar-Charged- Hydrophobiaty [91]. 

Pharmacophore profiles are defined that represent the set of all the pharmacophores found across the conformers of a series of conformers or series of molecules. Each pharmacophore added to the profile has to be unique. This profile will help in showing the spread of pharmacophores across the conformational space of a molecule or a series of molecules. No sum of the exhibited pharmacophores or normalization is done. There is no direct graphical representation of pharmacophore models. The pharmacophores can be saved to a file in CSV format that can be imported into a MySQL or Oracle database. This approach permits the use of standard SQL queries to extract common pharma-... 

Figure 13.4 Seven pharmacophore queries. The mesh balls represent H-bond acceptor sites, hydrophobic centers and aromatic centers. The solid white surfaces represent shape constrain. 

Successful virtual screening relies on an appropriate and correct description of the molecules being screened two particular sources of error are protonation state and tautomeric form. Oellien et have developed an exhaustive tautomer enumeration approach based on a set of 21 predefined transforms. The group do not attempt to say which tautomer is favoured, only that every form is represented. Databases built using all tautomers show better performance especially when screened with pharmacophore queries, as the bias introduced by most compound registration tools as to the correct tautomer is removed. 

Fig. 10 Definition of the pharmacophore query for the Kvl.5 channel. Representatives of the two compound series (A and B) were aligned(b) to derive the pharmacophore (a). (Copyright is permitted by Elsevier).

Researchers at Aventis described a pharmacophore for Kvl.5 channel blockers consisting of three hydrophobic centers in a triangular arrangement. First observed in an earlier active series [73], the pharmacophore was validated using a database of 423 Kvl.5 blockers [74], The query was able to retrieve 58% of the known actives. A pharmacophore search of the corporate collection identified 27 clusters containing 1975 compounds. Screening the representatives of 18 clusters led to the discovery of an anthranilamide hit molecule (Kvl.5 IC50 = 5.6pM in Xenopus oocytes) [74],... 

The pharmacophore models produced are then used as the query for 3D searching techniques, which try to find from a large database of potential lead compounds those that meet the geometrical requirements of the model. Compounds that are found to contain the correct arrangement are called hits and are candidates for screening. These hits differ from the substructure or 2D similarity hits in that the backbone of the structure may be quite different from that of the original lead compound and often represents an important new area of chemistry to be explored. 

---