References

Engel, eds., Chemoinformatics A Textbook, Wiley-VCH, Weinheim, 2003. 2. J. Gasteiger, ed., Handbook of Chemoinformatics From Data to Knowledge., Wiley-VCH, Weinheim, 2003. 3. T. Engel, J. Chem. Inf. Model., 2006, 46, 2267-2277. 4. W. L. Chen, J. Chem. Inf. Model., 2006, 46, 2230-2255. 5. N. Brown, Computing Surveys, 2006. 6. G. Harper, J. Bradshaw, J. C. Gittins, D. V. S. Green and A. R. Leach, J. Chem. Inf. Comput. Sci, 2001, 41, 1295-1300. 7. I. I. Baskin, M. I....

Topological Indexbased Pharmacophores

Arguments in favor of QSAR studies systematically included the claim that the models may help in elucidating ligand binding mechanisms. Unfortunately, recent results reveal that this is unlikely to be the case, unless draconian standards in terms of training set diversity and size are being fulfilled. Even in the presence of enlarged data sets, and benefiting from inclusion of a vast majority of published actives and thousands of experimentally certified inactives That information may be...

Pharmacophore Constraints Used in Docking

One of the most important applications for docking methods is the correct prediction of a possible binding pose, an ability that has been exhaustively studied for various docking programs, and ways to optimize the assessment of binding pose quality have been discussed.136,206 208 Sometimes, the available data on a given binding pocket and a corresponding set of known ligands may provide a very good reasoning for the assumption as to where exactly a certain part of the ligand is located inside...

Machinelearning of Topological Pharmacophores from Fingerprints

Machine learning can be used to select or weigh specific pharmacophore elements from a fingerprint, to improve the predictive power of the model, in letting it focus on the actually important patterns in the ligands and ignore unbound ligand moieties pending out into solvent (recall discussion in Section 2.3.3). An intermediate between similarity searching and descriptor-selection based QSAR, self-organizing maps68 (SOM), will be mentioned first. SOMs try to classify a population of individuals...

Antitarget and Admet Screening Using Pharmacophores

In the search for an optimal lead compound, not only activity on the desired target but also activity on other targets should be considered. Sometimes, multi-protein-targeting is desired, e.g., for multi-kinase inhibitors in the oncology field.165 However, also activity on side effect-related targets - so-called antitargets -is to be investigated. Compounds with no or very low affinity to proteins related to cardiovascular, cytotoxic, or metabolic effects are more likely to pass subsequent...

Structurebased Pharmacophore Modeling

During the last decade, structure-based methods (also called target-based or direct approaches), which rely on the availability of structural data of the target, have gained significant interest, since the number of experimentally determined three-dimensional structure of targets has grown constantly. Information on the target structure is preferentially taken from experimental investigations, but can also be taken from homology modeling.90 Today, the largest database of X-ray and NMR...

Topological Fingerprints

Topological fingerprints consider the connection table of a target molecule and ignore the respective atom coordinates. The rapid calculation and efficiency of such fingerprint methods implicate their wide acceptance in industry.117 An overview of similarity-based 2D fingerprint methods and their performance during VS is provided in ref. 176. Structurally similar molecules will often bind to the same group of proteins. While this hypothesis may be violated in specific cases - a small change in...

Topological Pharmacophore Triplets

To generate topological triplets, a basis set of reference pharmacophore triplets is chosen, enumerating all possible combinations of pharmacophore features of the corners, times all the considered integer edge lengths obeying triangle inequalities, within a finite range Emin, Emax . A basic example for such fingerprints are Typed Graph Triangles (TGT),44 a binary fingerprint monitoring the presence absence of each of the considered pharmacophore triangles. A series of improvements,48 such as...

Topological Pharmacophores from 2DAligments

Generating a 2D-aligment, the precursor step of 2D pharmacophore elucidation consists in first establishing, for each atom of the aligned molecule, a list of possible matches34 (equivalent atoms) in the target compound (subsequently referred as the template). Unless the aligned compound is a close analogue of the template, this problem is not trivial, for certain atoms may not have any appropriate equivalents in the partner compound. Putatively matching atom pairs are vertices of similar...

Current Superpositioning Techniques for Aligning 3D Pharmacophores and Molecules

In the broad field of possible pharmacophoric alignment techniques one can distinguish between either point- or property-based approaches.62 With point-based approaches, atoms or chemical feature point distances are minimized, while property-based approaches generate a pharmacophore by assessing the molecular interaction potential (MIP) similarity, based on Goodford's GRID63 method, to generate alignments. Programs representing both approaches have been applied for the generation of...

Application of Pharmacophore Models in Virtual Screening

As shown in the previous section, there have been an impressive number of new approaches and tools appearing in the field of structure- and ligand-based pharmacophore modeling. All of them address slightly different issues and therefore a combination of several methods, as shown in different application examples, can significantly enhance the chances of success. In our study on the discovery of novel ligands for the sigma-1 receptor and related proteins we were able to find compounds binding to...

Validation of Models for Virtual Screening

A useful pharmacophore model must be able to correctly classify - and in the case of quantitative models, correctly predict the affinity of - a so-called test or validation set, which consists of active compounds that were not used during the generation of the model. Often the ability of the model to retrieve known actives from a larger drug-like database is also assayed. It has been confirmed by several studies that the characteristics of the known inactives or decoys chosen for VS assessments...

Atomcentered Fragments

Atom-Centered Fragments (ACF) consist of a single central atom surrounded by one or several shells of atoms separated from the central one by the same topological distance. This type of structural fragments was introduced in the early 1950s by Tatevskii,27'28'116 119 and then by Benson31 to predict some physicochemical properties of organic compounds in the framework of additive schemes. ACF fragments containing only one shell of atoms around the central one (i.e., atom-centered neighborhoods...

Similarity Search

The notion of molecular similarity (or chemical similarity) is one of the most useful and at the same time one of the most contradictory concepts in chemoinformatics.247,248 The concept of molecular similarity plays an important role in many modern approaches to predicting the properties of chemical compounds, designing chemicals with a predefined set of properties and, especially, in conducting drug design studies by screening large databases containing structures of available (or potentially...

Summary and Conclusion

The pharmacophore concept is a successful and well-known approach - both ligand- and structure-based - for drug design as well as for VS. Several methods for describing pharmacophores have been established, showing significant differences and capabilities in the way in which they describe chemical features as building blocks for pharmacophores. It is important that the chemical feature representation used reflects the interactions that are relevant for the target being represented, and some...