QSAR

1       Introduction to Cerius2 Activity Prediction

Cerius² activity prediction is provided within an integrated set of software modules used to generate, manipulate, visualize, and analyze molecular structures, their conformers, and associated properties. These tasks, for instance, are typical of the drug discovery process. C²·QSAR+, in combination with other Cerius² modules, can be used to generate and analyze structure-activity relationships of potential lead compounds and to visualize and analyze receptor-pharmacophore interaction patterns.

Cerius² provides extensive property estimation and computation functions. It gives you the ability to derive pharmacophoric and receptor hypotheses and to predict pharmacological activity of known or novel compounds. All capabilities are packaged in modular form. You acquire and use only the Cerius² tools and functions that best fit your needs.

This chapter introduces you to activity and property prediction in Cerius²: an overview, the components, and primary modules. The chapter also includes information about starting Cerius² software.

Understanding Cerius2 activity prediction

The drug discovery process involves teams drawn from many disciplines. For computers and computer software to contribute significantly to the process, applications software must be designed so that information flows from discipline to discipline in such a way that it can be understood and used by everyone.

All Cerius² modules used for activity prediction are based on the common control and presentation platform of the Cerius² Visualizer. The Cerius² Visualizer is an integrated collection of tools for session management and atomistic modeling that is the core of the Cerius² modeling environment. These tools perform functions associated with model management, 3D graphical representation, session logging, environment customization including tables and graphs, and visualization of molecular models.

The activity prediction modules plug into the Visualizer and use the same graphical user interface. Some of these modules also are used by other Cerius² products. These include Tables & Graphs, Conformers, Open Force Field Setup (OFF Setup), Open Force Field Methods (OFF Methods), Quantum Mechanics, Merck molecular forcefield (MMFF), Field Calculation, and Builders. Additional builder modules are also available to add to your workbench to supplement the basic model sketching capabilities provided by the Cerius² Visualizer.

These are the Cerius² software modules specific to activity prediction:

• QSAR+ , which generates quantitative structure-activity relationship models in both basic default and customizable modes. It calculates 2D and 3D spatial, electronic, fragment, topological, thermodynamic, conformational, and shape properties (descriptors), and statistically analyzes relationships between molecular structures and the descriptors to provide correlations for predicting biological activity. More than 100 relevant descriptors are included, and new descriptors can be added.
• Molecular Field Analysis (MFA), which quantifies the interaction energy between a probe molecule and a set of aligned target molecules in a QSAR. Interaction energies measured and analyzed for a set of 3D structures can be useful in establishing QSARs.
• Genetic Function Approximation (GFA), which is a statistical analysis method that generates multiple QSAR models. Usually, this population of models contains many models comparable or superior to the single model generated with standard regression analysis. The multiple models are created by evolving random initial models using a genetic algorithm. The default is to build linear models, but other options, including higher order polynomials, splines, or other non-linear functions, also can be built. A method that combines genetic function approximation and partial least squares, G/PLS, is also available.
• Molecular Shape Analysis (MSA), which extends QSAR operations for performing 3D QSAR studies. This technique generates quantitative measurements of molecular shape properties as part of QSAR analysis.
• Diversity, which provides tools to build combinatorial libraries based on scaffold-plus-R-groups methods, and to optimize and visualize the diversity of combinatorial libraries.
• Alignment, which provides tools to superimpose molecules to satisfy various alignment conditions. These tools permit alignment of molecules using least square fitting with atom equivalencies specified either by automatic atom matching algorithms or by manual atom matching. In addition to rigid body superpositioning, the module provides tools for flexibly aligning one molecule over another using a fit optimizer algorithm.
• Interfaces for Catalyst ConFirm and Catalyst HipHop, which access Catalyst applications that provide tools to generate pharmacophoric hypotheses. The hypotheses are generated by first generating conformations for a set of study molecules and then using the conformations to find and align chemically important functional groups common to the molecules in the study set. The DRUG DISCOVERY card deck includes interfaces to both applications.
• Receptor, which provides a 3D visual environment for receptor hypothesis exploration. The module creates receptor surface models using information generated from the overlay of active compounds. The receptor models can be used to evaluate new compounds and evaluate conformations and constraints on compounds in the receptor site.
• Database Query, which provides access to the search facilities of Catalyst and ISIS to search molecular databases, retrieve, examine, and save structures that fit your search criteria.

Catalyst is an MSI product that performs hypothesis generation, database searches, and activity information. The database search functions, Cat/Info and CatShape, are accessed through Cerius² interfaces.

Before you begin

Before you use Cerius² for activity prediction, you must have installed on your system a properly licensed copy of the Cerius² software containing the modules you have purchased. If you have any questions about your system setup or your software license, please talk to your system administrator.

You should become familiar with the Cerius² environment. The Cerius² Modeling Environment introduces you to the Cerius² interface and to the Visualizer tools. The book also provides detailed information about standard elements of the modeling environment. Cerius² software runs on workstations using the UNIX operating system.

To start the C² -modules

Depending on the configuration that was chosen during Cerius² installation, you start the software using one of the following two commands in a UNIX shell window. You must be in a directory where have write access.

If Cerius² was configured to display DDW card decks, enter:

>	cerius2

>	cerius2 -x ddw 

The main Visualizer control panel contains the DDW card decks, including:

• DRUG DISCOVERY
• QSAR
• HYPOTHESIS MODELS
• DATABASES
• CONFORMATIONS
• COMBI-CHEM I and COMBI-CHEM II
• OFF SETUP and OFF METHODS
• TABLES & GRAPHS

Accessing C2 modules

If the full complement of DDW modules is installed, you can see the card decks described below on the main Visualizer control panel by selecting them from the deck selector popup.

DRUG DISCOVERY deck

The DRUG DISCOVERY card deck includes the QSAR, MODEL RECEPTOR, ALIGN MOLECULES, and QUERY DATABASE cards.

This deck includes tools to generate and apply QSARs. It also includes tools to align molecules, to generate and evaluate receptor models, and to construct molecular queries for searching databases and retrieving structures. For information on the QSAR menu selections, see the chapters later in this book. For information on ALIGN MOLECULES, MODEL RECEPTOR, and QUERY DATABASE, see Cerius² Hypothesis and Receptor Models.

QSAR deck

The QSAR card deck includes the QSAR, FIELD ANALYSIS (MFA), GENETIC ANALYSIS (GFA), and SHAPE ANALYSIS (MSA), and ORACLE READ cards. This deck includes tools to generate and work with QSARs, customize processes and parameters, execute molecular field and molecular shape analyses, and configure genetic function approximation analysis.

BUILDERS 1 deck

The BUILDERS 1 card deck allows you to construct various types of structures. These Cerius² builder modules are described in the Builders manual.

BUILDERS 2 deck

The BUILDERS 2 card deck contains a single card, ANALOG BUILDER.

The Analog Builder includes tools to generate structural analogs, to save structures (fragments) that you build, and to import MDL RG files.

COMBI-CHEM decks

The COMBI-CHEM I and COMBI-CHEM II card decks include tools to build and analyze combinatorial libraries. The cards in the COMBI-CHEM I deck are:

• LIBRARY CONSTRUCTION
• LIBRARY ANALYSIS
• LIBRARY COMPARISON
• BINARY DATA FILES
• ADVANCED BINNING

• SMILES
• LIB ENGINE
• QUERY DATABASE
• CONFIRM
• HIPHOP

The CONFORMATIONS card deck includes the CONFORMER SEARCH and CONFORMER ANALYSIS cards.

This deck includes tools to generate conformations using conformational search techniques, and to group and analyze the conformations using conformational analysis tools. For information on the menu selections in this deck, see Cerius2 Conformational Search and Analysis.

DATABASES deck

The DATABASES card deck includes the MDL INTERFACE, CATALYST INTERFACE, CATSHAPE INTERFACE, and ORACLE READ cards.

This deck includes tools to construct molecular queries for database searches and structure retrieval using either Catalyst or MDL ISIS and the databases associated with their search facilities.

FIELD CALCULATION deck

The FIELD CALCULATION card deck includes two cards, FIELD GENERATION and FIELD VISUALIZATION. These cards give access to tools to control the creation and display of probe surfaces.

HYPOTHESIS MODELS deck

The HYPOTHESIS MODELS card deck includes the CONFIRM, HIPHOP, and MODEL RECEPTOR cards.

MMFF deck

The MMFF card deck includes a single MMFF card, which enables you to control the Merck molecular forcefield.

OFF SETUP deck

The OFF SETUP card deck includes OPEN FORCE FIELD, CHARGES, and FORCE FIELD EDITOR.

This deck includes tools to load, edit, and apply forcefield parameters, including atom types, energy terms and expressions, charges, and geometry constraints. For information on the menu selections in this deck, see the Cerius² Simulation Tools.

OFF METHODS deck

The OFF METHODS card deck includes MINIMIZER, DYNAMICS SIMULATION, ANALYSIS, CONFORMER SEARCH, and CONFORMER ANALYSIS.

This deck includes tools for performing energy minimization, dynamics simulations, and analysis. For information on the menu selections in this deck, see Cerius² Simulation Tools.

PROTEIN TOOLS deck

The PROTEIN TOOLS card deck includes the ACTIVE SITE VIEWER and PROTEIN DISPLAY cards. The ACTIVE SITE VIEWER card provides tools to view and analyze both the structure and the environment of bound ligands and active sites.

QUANTUM decks

There are two quantum mechanics decks, which contain the DMOL3, MOPAC, GAUSSIAN, ADF, and ZINDO cards (QUANTUM 1) and the CASTEP, FASTSTRUCTURE, and ESOCS cards (QUANTUM 2).

These cards include tools for running externally provided programs.

TABLES & GRAPHS deck

The TABLES & GRAPHS card deck includes the TABLES, GRAPHS, and ISOSURFACES cards.

This deck includes tools to select, generate, edit, and file various types of graphs from a gallery of selections and datasets. The TABLES and GRAPHS cards includes tools to select, generate, edit, file, and manage tables and grids and the data in them. The ISOSURFACES card includes tools to generate surfaces from grid data. For information on the menu selections in this deck, see the Cerius² Modeling Environment.