QSAR

Contents

Release 4.5, June 2000

How To Use This Book

Preparing to work

How to find information

Using other Cerius2 books

Typographical conventions

1. Introduction to Cerius2 Activity Prediction

Understanding Cerius2 activity prediction

Working with activity prediction modules

Accessing C2 modules

2. QSAR+ QuickStart

Understanding the QSAR generation process

Using QSAR+

Creating a training set

Entering biological activity data

Calculating descriptors

Setting dependent and independent variables and exploring the data

Generating a QSAR equation

Analyzing the QSAR equations

Saving the QSAR equations

Predicting activity of new molecules

Saving the study

Summary

3. Theory: Statistical Methods

Principal components analysis (PCA)

Cluster analysis

Jarvis-Patrick clustering

Variable-length Jarvis-Patrick clustering

Relocation clustering

Hierarchical cluster analysis (HCA)

Step 1

Step 2

Regression methods

Simple linear regression (simple)

Multiple linear regression (linear)

Stepwise multiple linear regression (stepwise)

Principal components regression (PCR)

Partial least squares (PLS)

Genetic Function Approximation (GFA)

Genetic partial least squares (G/PLS)

Validation methods

Cross-validation

Randomization test

Evaluating QSAR equations

4. Theory: QSAR+ Descriptors

Fragment constants descriptors

Sm, Sp

F, R

pi

HA

HB

MR

Sterimol-L

Sterimol-B1 through B4

Sterimol-B5

Conformational descriptors

Electronic descriptors

Sum of atomic polarizabilities (Apol)

Dipole moment (Dipole)

Highest occupied molecular orbital energy (HOMO)

Lowest unoccupied molecular orbital energy (LUMO)

Superdelocalizability (Sr)

Receptor descriptors

IntraEnergy

InterEnergy

InterEleEnergy

InterVDWEnergy

MinIntraEnergy

StrainEnergy

Quantum mechanical descriptors

Graph-theoretic descriptors

Topological descriptors

Wiener index (W)

Zagreb index (Zagreb)

Hosoya index (Z)

Kier and Hall molecular connectivity index (c)

Example of the molecular connectivity index

Order zero

Order zero indices, CHI-0 Order one

Order one index Second-order

Second-order indices Third-order

Third-order indices Fourth-order

Fourth-order indices

Kier & Hall valence-modified connectivity index (cv)

Kier & Hall subgraph count index (SC)

Example of the Kier & Hall subgraph count index

Zeroeth-order indices

First-order indices

Second-order index

Third-order indices

Kier's shape indices (kn (n = 1, 2, 3))

Kier's alpha-modified shape indices (kan (n = 1, 2, 3))

Molecular flexibility index (f)

Balaban indices (JX and JY)

Information-content descriptors

Information of atomic composition index (IAC-mean, IAC-total)

Information indices based on the A-matrix

Information indices based on the D-matrix

Information indices based on the E-matrix and the ED-matrix

Multigraph information content indices (IC, BIC, CIC, SIC)

Molecular shape analysis (MSA) descriptors

Common overlap steric volume (COSV)

Difference volume (DIFFV)

Common overlap volume ratio (Fo)

Non-common overlap steric volume (NCOSV)

RMS to shape reference (ShapeRMS)

Volume of shape reference (SRVol)

Spatial descriptors

Shadow indices

Jurs descriptors based on partial charges mapped on surface area

Molecular surface area (Area)

Radius of gyration

Density (Density)

Principal moment of inertia (PMI)

Molecular volume (Vm)

Structural descriptors

Number of rotatable bonds (Rotlbonds)

Thermodynamic descriptors

AlogP, AlogP98, and molar refractivity (MolRef)

Desolvation free energy for water (Fh2o) and octanol (Foct)

Heat of formation (Hf)

pKa descriptors (ACD Labs)

Molecular field analysis (MFA) descriptors

Receptor surface analysis (RSA) descriptors

5. Working with the Study Table

Overview of the study table

Study table components

Study table menubar

File pulldown

Edit pulldown

Molecules pulldown

Descriptors pulldown

Variables pulldown

Tools/Table menu item

Tools/Graphics menu item

Tools/Statistical menu item

Other Tools menu items

Preferences pulldown

Using study table shortcuts

Basic study table operations

Displaying the current study table

Saving your work

Accessing a new, blank study table

Reloading an existing study table

Opening other table files

Exporting a study table

6. Working with Molecules

Loading molecules into Cerius2

Adding molecules to the study table

Setting molecule-processing preferences

Setting preferences for charges, minimization and conformations

Loading molecules directly from SD files

Special memory-saving options

Recovering deleted molecules

Loading molecules directly from Daylight SMILES files

Special memory-saving options

Recovering deleted molecules

SMARTS table derivations

Exporting molecules to SD files

Managing conformations

Displaying conformation information

Displaying contingent descriptors

7. Working with Descriptors

Default descriptors sets

QSAR defaults descriptor set

Managing descriptors

Using the default descriptors

Selecting descriptors

Setting descriptors preferences

Daylight descriptors preferences

Information-content descriptor preferences

Receptor descriptor preferences

Receptor surface analysis (RSA) preferences

Spatial preferences

Defining hydrogen-bond acceptors and donors and rotatable bonds

Thermodynamic descriptors preferences

Topological descriptors preferences

Adding descriptors to the study table

Using ISIS keys and Daylight fingerprints

ISIS keys

Daylight fingerprints

Using receptor surface analysis descriptor

Using pKa descriptors

Installing pKa

Adding pKa descriptors to the study table

What do the pKa column names mean?

Editing a descriptor database

Opening a descriptor database

Identifying default descriptors

Adding a descriptor to the default set

Removing a descriptor from the default set

Creating new descriptors

Modifying descriptors

Controlling the descriptor display format

Creating new descriptor categories

Saving a descriptor database

8. Working with Fragment Constants

Selecting fragment constants

Identifying fragment positions in study molecules

Renaming fragments

Core searching

Editing the fragment constants database

9. Performing Molecular Field Analysis

Accessing molecular field analysis

Creating a field

Changing field parameters

Setting MFA preferences

Managing independent variables

Managing fields

Generating a QSAR using field data

Predicting biological activity

10. Performing Molecular Shape Analysis

Accessing molecular shape analysis

Overview of molecular shape analysis

1. Generating conformations

Setting conformation generation preferences

Accessing the QSAR Conformation Generation control panel

Selecting a conformation generation method

Applying an energy cutoff

Specifying the number of conformers

Generating conformations

2. Hypothesizing an active conformer

Accessing the Active Conformation control panel

Selecting the active conformer

Displaying the active conformer

3. Identifying a shape reference compound

Accessing the Shape Reference control panel

Selecting a shape reference compound

Displaying the selected shape reference compound

4. Aligning molecules

Accessing the Shape Reference control panel

Aligning models

Aligning models by MCS method

Aligning models by CSS method

Removing alignment information

5. Measure molecular shape commonality

6. Determining other molecular features

7. Generating a trial QSAR

Accessing the Select Conformers control panel

Selecting conformers

Generating a trial QSAR equation

11. Working with Variables and Observations

Identifying variables using study table icons

Identifying variables using the Select Variables control panel

Selecting variables

Identifying variables

Changing the type of a variable

Changing the QSAR+ default variable settings

Resetting variables

Selecting observations

12. Working with Statistics

Selecting a statistical method

Genetic function approximation

Genetic partial least squares (G/PLS)

Multiple linear regression

Partial least squares

Principal components analysis

Principal components regression

Simple linear regression

Stepwise multiple linear regression

Presenting QSAR statistical results

Analysis of variance (ANOVA) table

Beta coefficient table

Equation viewer

Plots

Validating QSAR equations and data

Setting the default validation option

Using other validation procedures

Working with outliers

Displaying statistical information for exploratory data analysis

Displaying a correlation matrix

Displaying a descriptive statistics table

Displaying rune plots

13. Genetic Function Approximation

Overview of genetic function approximation

Using genetic partial least squares

Starting a genetic analysis

Performing a genetic analysis

1. Starting the analysis

2. Building the initial population

3. Evolving the population

4. Reviewing the evolved equations

5. Using the equations

Working with the current equation population

Continuing the evolution of the current population

Randomizing the current population

Setting genetic analysis preferences

Selecting equation term types

Specifying mutation probabilities

Specifying other genetic analysis preferences

Establishing the population size

Setting the smoothing parameter d

Setting the number of equation terms

Setting the length of the equation

Setting the regression method

Using genetic partial least squares

Running a G/PLS calculation

Setting G/PLS preferences

14. Using the Equation Viewer

Opening the equation viewer

Selecting equations

Deleting equations

Plotting equations

Renaming equation sets

Saving QSAR equations

Opening QSAR equations

Deleting a QSAR equation

Labelling 3D QSAR equations

15. Classification Structure-Activity Relationship (CSAR)

Recursive partitioning

Controls in the Interacting with Plot section of the Recursive Partitioning Preferences control panel

Select

Information

List Members

Model Construction Options in the control panel

Weighting options

Scoring splits

Pruning

Samples per node

Knot limits

Maximum tree depth

Membership list printout

Study Table Options section of control panel

Class probability columns

Crossvalidation Test section of control panel

Interpreting the results

Tutorial: Deriving a decision tree

A. References

B. Tutorial: Building a QSAR equation

Before you begin

Entering the molecules in the training set

Entering molecular descriptors

Load the molecules into the QSAR study table

Entering biological activity data

Exploring the data

Generate a QSAR equation

Analyzing the QSAR equation

Saving the QSAR equations

Predicting the activity of new molecules

Saving the study

Summary

C. Tutorial: Managing SD files in QSAR+

Before you begin

Lesson 1: Exporting an SD file

Lesson 2: Importing an SD file

Summary

D. Tutorial: Principal Component Analysis

Before you begin

Solving the problem

References

E. Tutorial: Cluster Analysis

Before you begin

Solving the problem

Reviewing the solution

References and related material

F. Tutorial: Fragment Constants Tutorial

Introduction

Before you begin

Generating study molecules from a core model and the default database

Generating study molecules from a different fragment library

A complete example using existing models

Reference

G. Tutorial: CSAR Tutorial

Recursive partitioning

Using CSAR with binary data files

Index