Imaginative chemical modelling facilitating drug discovery

pdf version (289KB)

Hugo Kubinyi, Professor of Pharmaceutical Chemistry at the University of Heidelberg, Germany, former Chair of the QSAR and Modelling Society and former head of Drug Design at BASF AG, Germany

The QSAR Builder computational environment enables researchers to generate and evaluate QSAR and QSPR models and develop qualitative structure activity relationships (C-SAR).

QSAR Builder implements fragmental methods in the analysis of any property or activity using clearly defined rule-based routines or user-developed custom fragmention sequences. Fragmental methods can also be effectively applied to smaller data sets when investigating congeneric series related to specific ligand/pharmacophore interactions or bioactivity.

Benefits

• Unified workspace streamlines the modelling process: filter data sets, apply fragmentation methods, view matrices, run statistical routines and analyze results using a single integrated application.
• Rapid hypothesis testing: edit and save multiple molecule lists, test different sets of variables and apply diverse statistical methods within a single project file.
• Improve communication among researchers working in different areas: all results are interpreted in terms of chemical structures.

Features

Workspace
QSAR Builder provides a logical, project-driven workspace employing object-based tools and menu-driven commands. Members of a discovery team can visualize strategy development and optimisation methods. Files can be saved, ported and modified as a project develops.

Fragmentation Tools
Five customizable fragmentation methods enable flexible analysis and structural descriptor derivation: Atom-based (constituent atoms), Hansch-Leo (isolating carbons, functional groups and interactions), Free-Wilson (skeletons and radicals), Klopman-type (atom chains), and Scaffold-based.

Descriptors
Phys-chem algorithms provide calculated descriptor values. This set of descriptors is strongly linked to drug optimization: Excess molar refraction (N~700), H-bond acidity, H-donor capability (N~700), H-bond basicity, H-acceptor capability (N~700), McGowan's Volume [Abraham, 1987], Hexadecane/gas partition coefficient (N~700), LogP partition coefficient (N~9,800), TPSA - topological polar surface area [Ertl, 2000].

Database
Importing of SMILES, SDF, RDF, Mol and tab-delimited files with OLE linking to ChemDraw, ISIS/Draw. Filter and search by structure, substructure and field. Create, save, edit and merge molecule lists.

Data Matrix
Matrices display the structures and fragments being investigated in rows. Columns show descriptor and increment information generated during optimization studies. CPU-friendly matrix objects allow for accelerated statistical analysis and high throughput calculations for data sets up to 100,000 x 50,000.

Expression Editor
Flexible and understandable menu-driven query builder. Provides tools for rapid assembly and application of database filters, screens and calculators.

Statistical Methods
These methods have been specially adapted to the unique requirements of medicinal chemists:

QSAR Builder's workspace unifies data filtering and chemical spreadsheet components as well as tools for descriptor calculation, fragmental and statistical analysis.