QSAR

8       Working with descriptors

A descriptor is any one of a number of molecular properties that QSAR+ can calculate and use in determining new QSAR relationships. QSAR+ provides over 100 different descriptors in a variety of categories:

• Spatial

• Electronic

• Thermodynamic

• Conformational

• Topological

• Information-content

• Quantum mechanical

• Structural

• Descriptors based on fragment constants

• Descriptors based on receptor surface models

• Descriptors based on molecular field analysis (MFA)

• Descriptors based on molecular shape analysis (MSA)

To meet your requirements for building QSAR equations, QSAR+ enables you to work with descriptors in a variety of ways.

Managing descriptors

You manage descriptors by using the several control panels (described later in this chapter). Descriptor management includes activities such as identifying the descriptors with which you want to work, displaying and selecting only descriptors in a specific class, specifying preferences for the various descriptors, and adding descriptors to the study table.

Editing the descriptor database

When QSAR+ is installed, you can access a descriptor database that contains the equations used to calculate molecular descriptors. You can edit this database to modify the supplied descriptors, create new descriptors, specify which descriptors should be considered default descriptors, create new descriptor categories, and control the format in which the results of descriptor calculations are displayed in the study table.

The following activities related to working with descriptors are included in this chapter:

Default descriptors sets in the following section.
Managing descriptors on page 148.
Using receptor surface analysis descriptor on page 155.
Rule of five on page 162.

In Start-up and Configuration:

Daylight setup
Oracle setup
MDL ISIS setup

Using molecular diversity descriptors
Statistical analyses and data-mining techniques

Graph-theoretic descriptors
Information-theoretic descriptors
Descriptors based on projections of the molecular surface (shadow indices)
Descriptors based on partial charges mapped on surface area
2D and 3D fingerprints metrics

Visualization of compounds in descriptor space
Principal component analysis
Factor analysis
Cluster analysis
Multidimensional scaling

You can see the descriptors in each set by selecting Descriptors/Databases from the study table menu bar. This opens the Descriptor Database control panel, which contains a list of descriptors.

The message at the top of the Descriptor Database control panel identifies the current default set.

QSAR defaults descriptor set

Table 2. QSAR default descriptors

Conformational
EPenalty	Conformational energy penalty.
LowEne	Lowest energy conformer.
Energy	Energy.
Electronic
Charge	Sum of partial charges.
Fcharge	Sum of formal charges.
Apol	Sum of atomic polarizabilities.
Dipole	Dipole moment.
HOMO	Highest occupied molecular orbital.
LUMO	Lowest unoccupied molecular orbital.
Sr	Superdelocalizability.
Information
InfoContent	Graph-theoretical Information-content indices.
Molecular shape analysis (MSA)
DIFFV	Difference volume.
Fo	Common overlap volume (ratio).
NCOSV	Non-common overlap steric volume.
ShapeRMS	Rms to shape reference.
COSV	Common overlap steric volume.
SRVol	Shape reference volume.
Quantum mechanical
LUMO_MOPAC	Lowest unoccupied molecular orbital from MOPAC.
DIPOLE_MOPAC	Dipole moment from MOPAC.
HF_MOPAC	Heat of formation from MOPAC.
HOMO_MOPAC	Highest occupied molecular orbital from MOPAC.
Receptor
Receptor_energies	Molecule-receptor interaction energies.
Receptor_RSA	Molecule-receptor points interaction energies.
Spatial
RadOfGyration	Radius of gyration.
Jurs descriptors	Jurs charged partial surface areas descriptors.
Shadow indices	Surface area projections descriptors.
Area	Molecular surface area.
Density	Density.
PMI	Principal moment of inertia.
Vm	Molecular volume.
Structural
MW	Molecular weight.
Rotlbonds	Number of rotatable bonds.
Hbond acceptor	Number of hydrogen-bond acceptor groups.
Hbond donor	Number of hydrogen-bond donor groups.
Chiral centers	Count of the number of chiral centers (R or S) present in a molecule.
Thermodynamic
AlogP	Ghose and Crippen logP.
AlogP98	Log of the partition coefficient, atom-type value.
Fh2o	Desolvation free energy for water.
Foct	Desolvation free energy for octanol.
Hf	Heat of formation.
MolRef	Ghose and Crippen molar refractivity.
Topological
Balaban	Balaban indices.
Kappa indices	Molecular shape kappa indices.
PHI	Molecular flexibility index.
SubgraphCount	Subgraph counts.
Chi indices	Kier & Hall chi connectivity indices.
Wiener	Wiener Index.
log Z	Logarithm of Hosoya index.
Zagreb	Zagreb index.

• Using the default descriptors

• Selecting descriptors

• Setting descriptors preferences

• Adding descriptors to the study table

To add the default descriptors set to the study table, select the Descriptors/Add Default menu item in the study table. This adds the current descriptors database to the study table. A button in the study table is also available to do this.

Selecting descriptors

The Descriptors control panel contains a list of the descriptors in the current descriptors database. These may be selected by clicking the descriptor name in the first column, for example, clicking EPenalty causes that row of the descriptor table to become highlighted, which means it will be added to the study table (see the next section for details). To unselect a descriptor, click any part of the table other than the first column, so that the highlight is turned off.

The Descriptors control panel contains controls that allow you to select groups of descriptors. The left popup controls whether the action that occurs when you click the associated action button is to Select, Deselect, or Display the selected descriptors. For example, if you want to select all the conformational descriptors, you can do so by choosing Select in the left popup and then setting the Descriptors in Family popup (far right) to Conformational. Now when you click the (unlabeled) action button (below ADD), the conformational descriptors are selected. To deselect them, change the Display popup to Deselect, then click the action button again.

If you find the display of all the descriptors at the same time distracting, you can display just the selected descriptors by setting the popup to Display.

Another way to select a subset of descriptors is to use the All/Default popup. To see the effect of this control, set the Descriptors in Family popup to Electronic, select Default from the All/Default popup, then click the action button.

Setting descriptors preferences

When the Descriptors in Family popup is set to Electronic, for example, the Preferences button is labelled Electronic. When you click this newly active pushbutton, a control panel appears, which allows you to customize certain aspects of the way the electronic descriptors are calculated. For example, if you decide that only the total dipole moment is needed, uncheck the XYZ Components checkbox. Now only the total dipole moment (calculated from atomic partial charges) is added to the study table.

Preferences for the calculation of other types of descriptors can be set in the same way.

Daylight descriptors preferences

Information-content descriptor preferences

If Edge-based is checked, the four buttons below apply to information indices based on the edge adjacency and edge distance matrices, specifically,

Edge adjacency/magnitude
Edge adjacency/equality
Edge distance/magnitude
Edge distance/equality

Vertex adjacency/magnitude
Vertex adjacency/equality
Vertex distance/magnitude
Vertex distance/equality

For a detailed explanation of this descriptor, see Chapter 5, Theory: QSAR+ descriptors.

Receptor descriptor preferences

You cannot add receptor descriptors to the study table until you have specified a receptor surface model. For information on this, see Using receptor surface analysis descriptor on page 155.

Spatial preferences

This control panel controls the calculation of spatial descriptors such as the moment of inertia about the principal axes of a molecule. For example, if you want the magnitude of the moment of inertia, but not its Cartesian components, uncheck the XYZ Components checkbox. See the Principal moment of inertia (PMI) section, page 101, for a theoretical explanation of the principal moment of inertia descriptor.

Jurs charged partial surface area parameters

The definition of polar atoms and the probe radius for the solvent-accessible surface area calculation can also be customized with the Spatial Descriptors control panel.

Polar atoms can be defined:

• From Partial Charges. All atoms with an absolute value of partial charge equal to or greater than a specified threshold are considered polar.

• From Atoms List. All hydrogen atoms attached to the elements listed in the associated entry box are also considered polar.

Checkbox	Toggles calculation of descriptors
Solvent Accessible Surface Area	SAS area descriptor Jurs-SASA
Partial Charged Surface Areas	Jurs-PPSA-1, Jurs-PNSA-1, Jurs-DPSA-1
Total Charge Weighted Surface Areas	Jurs-PPSA-2, Jurs-PNSA-2 and Jurs-DPSA-2
Atomic Charge Weighted Surface Areas	Jurs-PPSA-3, Jurs-PNSA-3, and Jurs-DPSA-4
Fractional Charged Partial Surface Areas	Jurs-FPSA-1, Jurs-FPSA-2, Jurs-FPSA-3, Jurs-FNSA-1, Jurs-FNSA-2, Jurs-FNSA-3
Surface Weighted Charged Partial Surface Areas	Jurs-WPSA-1, Jurs-WPSA-2, Jurs-WPSA-3, Jurs-WNSA-1, Jurs-WNSA-2, Jurs-WNSA-3
Relative Positive and Negative Charges	Jurs-RPCG, Jurs-RNCG, Jurs-RPCS, Jurs-RNCS
Relative Polar and Apolar Surface Areas	Jurs-TPSA, Jurs-TASA, Jurs-RPSA, and Jurs-RASA

Shadow indices

For an explanation of the shadow indices see the Shadow indices section on page 97 under Theory. The correlation between the Shadow Parameters checkboxes and the descriptor names is:

Checkbox	Toggles calculation of descriptors
Areas of Molecular Shadows	Shadow-XY, Shadow-XZ, and Shadow-YZ
Fractional Areas of Molecular Shadows	Shadow-XYfr, Shadow-XZfr, and Shadow-YZfr
Extents of Molecular Shadows	Shadow-nu, Shadow-Xleng, Shadow-Yleng, and Shadow-Zleng

Defining hydrogen-bond acceptors and donors and rotatable bonds

Open this control panel by setting the family popup in the Descriptors control panel to Structural and then selecting the Structural pushbutton.

Thermodynamic descriptors preferences

The 115 atom types defined in the calculation of AlogP98 are now available as descriptors. To calculate them, select the entry AlogP_atypes in the Thermodynamic family in the descriptor table. Each AlogP98 atom-type value represents the number of atoms of that type in the molecule. An additional atom type called Unkown_Type can also be added to the table, together with the other AlogP98 atom types. A value greater than zero for this descriptor indicates the presence of atoms that couldn't be classified as any of the defined AlogP98 atom types. The AlogP Atom Types control panel allows you to select the elements to be taken into account.

Open this control panel by setting the family popup in the Descriptors control panel to Thermodynamic and then selecting the Thermodynamic pushbutton.

Topological descriptors preferences

To change preferences for topological descriptors, set the family popup in the Descriptors control panel to Topological and select the Topological pushbutton. The correlation between the checkboxes in the Topological Descriptors control panel and the descriptors is:

Checkbox	Toggles calculation of descriptors
Unmodified	Molecular connectivity Indices CHI-0, CHI-1, and CHI-2
Valence-modified	Valence-modified connectivity index, a refinement which takes into account the atomic number and order of connected bonds.
Subgraph Order From and To	Range of allowable orders in subgraphs: 0 through M, where M is the number of edges in the graph.
Subgraph Type	Checkboxes Path, Cluster, Path/Cluster, and Ring specify the subgraph types used with the molecular and valence-modified connectivity indices.
Kier & Hall Kappa Shape Indices	Shapes of molecules in terms of the count of atoms (One), count of branchings (Two), and count of paths of length 3 (Three).
Subgraph Counts	Path, Cluster, Path/Cluster, and Ring subgraphs found in the model.
Balaban Indices	Characterize the shape of a molecule, which can take account of the covalent radii (JX) and electronegativity (JY) of the atoms of the model.

Adding descriptors to the study table

Using ISIS keys and Daylight fingerprints

• Calculate Isis keys for a SD file and save the results in an Isis keys file (the MDL environment should be properly defined fiirst).

• Import Isis keys from an Isis keys file into the study table. Molecule names read from the SD file are matched against names of molecules already present in the study table. For those that are found, the corresponding Isis keys are added to the appropriate rows; for any not found, new rows are added to the study table.

• Calculate Daylight fingerprints for a SMILES file and save the results in a TDT file (the Daylight environment should be properly defined first). The fingerprints will have length 1024. Molecule names, if they are included in the SMILES file, are saved in the TDT file with the tag MODEL_NAME.

• Import Daylight fingerprints from a TDT file into the study table. Molecule names read from the TDT file are matched against names of molecules already present in the study table. For those that are found, the correspondinf fingerprints are added to the appropriate rows; for any not found, new rows are added to the study table.

The first control panel (Receptor-Model Interactions) is concerned with addition of the receptor energy descriptors to the study table. To learn more about the receptor energy descriptors, see Receptor descriptors under Theory.

The second control panel (RSA Preferences) controls the addition of interaction energies at each vertex of the surface. You may add only the van der Waals (steric) component of the interaction energy or only the electrostatic component or both, by checking the VDW, ELE, and TOT (total) checkboxes.

A column is created in the study table for each point on the receptor surface model, containing the energy of interaction at that point between the surface and the molecule. For a large receptor surface model, this can be several thousands of columns if all points are added to the study table: too many for some of the statistical methods available. You can reduce the number of points added to the study table by using the Filter Surface Points popup.

Three main methods are available, based on selecting every n^th surface point or on adding points based on their variance or correlation. When you set the Filter Surface Points popup to the desired method, additional controls appear.

• Selection of every n^th surface point:

a.   Add all surface points

Add all the points on the surface of the receptor model to the study table.

b.   Add every Nth surface point

Add every n^th point on the surface of the receptor model to the study table. Typically this is a good place to start. Fill in the Every entry box with the frequency with which the surface is sampled.

• Selection based on variance:

The difference in energy at each surface point between each molecular model is used to filter input to the study table.

a.   Add points with variance higher than threshold

Those points with variance higher than the Variance Threshold are added to the study table.

b.   Add percentage of points with highest variance

The Percent is the percentage of highest-variance points to add.

• Selection based on correlation:

The square of the correlation between energy at each surface point and biological activity data in the study table (marked Independent Y) is used to filter the RSA input to the study table.

a.   Add points with correlation higher than threshold

Correlation^2 is used to filter out any columns that show lower correlation with the activity than the specified threshold.

b.   Add percentage of points with highest correlation^2

The Percent specifies the percentage of highest-correlation squared points to add.

Next, click the action button on the extreme left side of the Descriptors control panel (underneath the ADD button). This displays the receptor descriptors Receptor_energies and Receptor_RSA. To select the Receptor_RSA descriptor, click the cell containing the label Receptor_RSA. To add the receptor surface data to the study table, then click the ADD pushbutton. The receptor surface points are added to the study table.

These points may be displayed with the Manage Independent Columns control panel, which is accessed by selecting the Variables/Manage Independent menu item in the study table. Set the 3D-QSAR Labels popup to RSA and click the Label Independent Variables action button.

Surface points in the study table are displayed on the receptor surface model as a label, for example, TOT/123. The first part of the label refers to the type of energy term specified in the RSA Preferences control panel under Include Molecule-Surface Point Interaction Energies. The second part is the number of the surface point and is the same index as the Surface point index in the first column of the output of the Receptor List function.

Typically, the next stage is to calculate a QSAR that relates the receptor surface energy at each surface point to experimental activity data. For a guide to calculating QSARs, see Chapter 15, Using the equation viewer, and Chapter 3, QSAR+ QuickStart.

Using pKa descriptors

A unix pKa pathname 

Adding pKa descriptors to the study table

1.   Open the appropriate descriptor database

From the study table, open the Descriptor Database control panel by selecting the Descriptors/Databases menu item. In the Descriptor Database control panel, set the popup to the appropriate database and click the OPEN DATABASE pushbutton.

From the study table, open the Descriptors control panel by selecting the Descriptors/Select menu item.

2.   Set the pK_a descriptor preferences:

Set the family popup to ACD. Click the ACD pushbutton to open the ACD Descriptors control panel, which is used to set preferences for treating the pK_a data. Two types of pK_a descriptors are available: a count of pK_as for each model, and a list of pK_as.

To add a count of pK_as to the study table, check the List a count of pKas checkbox and specify the range within which pK_as are to be counted.

To add the pK_a values to the study table, check the lower List checkbox, specify whether the values should be listed from High to Low or from Low to High, specify the maximum number of pK_as to be listed, and specify a range or an upper or lower limit for pK_as to be listed.

3.   Add the pK_a descriptors to the study table

In the Descriptors control panel select the pKa row and click the ADD button. The pK_a descriptors are added to the study table, and the results are calculated for any entries already in the study table. For subsequent additions to the study table, the pK_a descriptors are calculated automatically.

A count of pK_a columns begins with the string n_pKa_. This is followed by the range of values being counted. For example, n_pKa_0.00_14.00 is a count of pK_as with values between 0.00 and 14.00.

A list of pK_a columns begins with the string pKa_. The first number tells which pK_a value among the selected pK_as is held in this column. The second number gives the maximum number of pK_as to be listed. The third number specifies whether the pK_as are listed from low to high (number = 0) or from high to low (number = 1), The fourth number specifies whether a range (number = 0) or a lower (number = 1) or upper (number = 2) bound is used to select the pK_as to list. If a range is used, it is followed by two numbers specifying the range. If a lower or upper bound is used, it is followed by the number specifying the bound. For example, pKa_1_2_0_2_14.00 is the lowest pK_a of a maximum of two pK_as under the bound of 14.00.

Using ADME descriptors

The panel is divided into three sections, one each for the ADME models (Egan et. al 2000). Each is described in the following sections.

Intestinal Absorption Model

First select a model type:

2D Model: model is based on Polar Surface Area (PSA) calculated from connectivity data only (2D structure)

3D Model: model is based on Polar Surface Area (PSA) calculated from 3D coordinates

2D and 3D Models: both 2D and 3D models are calculated

T-squared Values Check this to include a column with the T² values (distance from the center of the ellipse) used in the ADME absorption model calculation.

AlogP98 Values Check this to include a column with the AlogP98 values used in the ADME calculation.

PSA Values Check this to include a column with the polar surface area values used in the ADME calculation.

First select a model type:

2D Model: model is based on Polar Surface Area (PSA) calculated from connectivity data only (2D structure).

3D Model: model is based on Polar Surface Area (PSA) calculated from 3D coordinates.

2D and 3D Models: both 2D and 3D models are calculated.

BBB LeveValues Check this to include a column with the BBB Penetration level corresponding to the calculated Brain-Blood ratio.

Table 3. Key to BBB penetration levels

level	values	description
0	Very High	Brain-Blood ratio greater than 5:1
1	High	Brain-Blood ratio between 1:1 and 5:1
2	Medium	Brain-Blood ratio between 0.3:1 and 1:1
3	Low	Brain-Blood ratio less than 0.3:1
4	Undefined	Outside 99% confidence ellipse
5	Alogp98	Warning: molecules with one or more unknown Alogp98 types

AlogP98 Values Check this to include a column with the AlogP98 values used in the ADME calculation.

PSA Values Check this to include a column with the polar surface area values used in the ADME calculation.

Report Solubility Level Values: Check this to include a column of solubility levels corresponding to the logarithm of the water solubility.

Table 4. Key to water solubility levels

LogSw	Level	Description
< -8.0	0	extremely low solubility, lower than 95% of drugs
-8.0 to -6.0	1	very low solubility, at border line of 95% of drugs
-6.0 to -4.0	2	low solubility, at lower end of 95% of drugs
-4.0 to -2.0	3	good,slight soluble to soluble
-2.0 to 0.0	4	optimal solubility
> 0.0	5	very soluble, perhaps too soluble
1000	6	Warning: molecules with one or more unknown Alogp98 types

Rule of five

MW <= 500
Hbond acceptors <= 5
Hbond donors <= 10
logP <= 5

Set Hydrogen Bonds Preferences for Rule of 5 Click this to bring up the HBond Descriptors concrol panel allowing you to set preferences for Hbond donors and acceptors.

Report Number of Unknown AlogP98 Atom Types Check this to include a column with the number of unknown atom types for each molecule in the AlogP98 calculation.

Don't Calculate ADME Descriptors for molecules with Unknown AlogP98 Atom Types Check this to set ADME descriptors to undefined whenever the corresponding molecule has undefined AlogP98 atom types.

Analyzing ADME descriptors

In addition to displaying the molecular structure in the Cerius² models window, double-clicking the row name in the Study Table also displays the calculated ADME properties for the corresponding molecule, including Absorption, BBB Penetration and Solubility levels.

Intestinal Absorption Model

2D absorption model obtained using polar surface areas calculated from molecular connectivity.

3D absorption model obtained using the X, Y, Z coordinates to calculate polar surface areas.

The PLOT button generates a plot of PSA vs. AlogP98, such as the one shown below.

Two check boxes below specify the display of the 95% and 99% confidence limit ellipses obtained in the development of the model (Lipinski et al. 1997).

There are also options to display BBB Penetration model ellipses, which occupy a slightly different position in the plot. The Absorption level is calculated based on the position of each molecule in the PSA vs. AlogP98 plot:

Table 6. Key to water absorption levels

value	level	description
0	Good	Inside 95% Ellipse
1	Moderate	Inside 99% Ellipse
2	Poor	Inside box defined by PSA between 0 and 150 and AlogP98 between -2 and 7
3	Very Poor	Outside box
4	Undefined	Molecules with unknown AlogP98 atom types

By default the plot is centered on the good absorption areas (around the 95% and 99% ellipses) and the points are color-coded according to Absorption level.

Points can also be color-coded according to the number of violations to the Rule of 5. Additional controls allow the user to specify which points to include in the plot: points with Absorption level Greater or equal to, Equal to, or Smaller or equal to a specified level.

Molecules corresponding to points in the plot can be identified by using the Plot Pick Tool, which selects all the molecules within the specified PSA and AlogP98 distance of the click position and highlights them in the Study Table.

The PRINT button prints out the percentage of molecules in each Absorption level.

The SELECT button selects all the molecules in the Study Table which satisfy the specified Absorption levels.

The Absorption panel launched from the Binary Data File control panel is the same as the one described above, except the SELECT button is replaced by CREATE BDF FILE which creates a new BDF file with the molecules that satisfy the specified criteria.

BBB Penetration Model

The log₁₀ of the Brain-Blood ratio (logBB) is calculated only for molecules whose PSA and AlogP98 values lie within the 99% confidence limit ellipse corresponding to the BBB Penetration model. Points outside the ellipse are classified as undefined (probably poor penetration). For those points within the ellipse, a QSAR equation is used to predict logBB. The BBB Penetration levels are defined as follows (see Table 3).

Try pusing the PRINT button to generate sample output.

Water Solubility Model

Editing a descriptor database

You can modify an existing descriptor database or create a new one by editing the installed descriptor database provided in QSAR+, then saving the modified descriptor database under a new name.

This section describes the following activities related to editing a descriptor database:

• Opening a descriptor database

• Identifying default descriptors

• Creating new descriptors

• Modifying descriptors

• Controlling the descriptor display format

• Creating new descriptor categories

• Saving a descriptor database

Because the descriptor database is accessed as a Cerius2 table, you should be familiar with Cerius2 tables before performing any activities described in this section. For information about tables and basic table operations, see Cerius² Modeling Environment.

Opening a descriptor database

To open a descriptor d-atabase

If you have only a single database or if you want to use the currently selected database, select Descriptors/Databases in the study table or on the QSAR card. The Descriptor Database control panel appears.

If you have more than one descriptor database and want to change the selected database:

• Select Descriptors/Database in the study table or on the QSAR card. Select the QSAR, COMBICHEM, or QSPR descriptor set by choosing from the popup at the top of the Descriptor Database control panel.and click the OPEN DATABASE pushbutton.

or:

• If you want to open some other database file, set the popup to Other, which causes the Open Database control panel to appear when you click OPEN DATABASE. Use the file-selection tools to identify the descriptor database that contains the descriptors you want. Click OPEN.

The descriptor database table contains one row for each descriptor. Each row contains columns, some of which are described below (to see all columns, use the horizontal scroll bar).

• Descriptor name -- This column contains a name for each descriptor; for example, Rotbonds (rotatable bonds). For a list of descriptors that are part of QSAR+, see Chapter 5, Theory: QSAR+ descriptors.

• Family -- This column contains the family name for each descriptor. It is used for sorting and displaying groups of similar descriptors. The family name QSAR is automatically created when you save a QSAR equation in the descriptor database.

You can create your own family names (as described in Creating new descriptor categories on page 171), but usually most descriptors fit into one of the existing groups.

• Value -- This column contains a descriptor equation, made up of math and molecular operators. It defines the QSAR+ calculation used to generate the descriptor values.

• 3D -- If a descriptor is a 3D descriptor, the cell entry is Yes. If the descriptor is not 3D, the entry is No.

• Default -- If the descriptor is set as a default descriptor, the cell entry is Yes. If the descriptor is not a default descriptor, the entry is No.

• Format -- This column contains the numerical format for a calculated descriptor value: floating decimal (float), integer (integer), or scientific notation (scientific).

• Panel -- This column contains the name of the control panel (if any) accessed to edit descriptor properties.

• Token -- A command attached to the descriptor.

• Command -- Application that executes the attached command.

• Multiple -- Unused.

• Multiple_col -- If the descriptor creates multiple columns, this column contains Yes, otherwise it contains No.

You can change the set of default descriptors by editing the Default column.

Adding a descriptor to the default set

1.   Select the cell in the Default column for that descriptor.

2.   Clear the edit window and enter 1.

3.   Press <Return> or click any other cell in the table.

The Default cell now displays Yes.

1.   Select a cell in the Default column.

2.   Press <Return> or clear the edit window and enter 0.

3.   Click any other cell in the table.

The Default cell now displays No.

• Any valid math operator.

• A molecular operator that is supplied with the Cerius2·Visualizer.

• A molecular operator supplied with any separately licensed Cerius2 module that you have installed (for example, QSAR+).

1.   Insert a new row in the Descriptor Database table using the Insert tool.

2.   In the Family column of the new row, enter a family name.

Most descriptors can be categorized into one of the existing families, usually spatial, electronic, or thermodynamic. For example, if you want to create a descriptor that counts halogen atoms in a structure, enter spatial.

3.   Enter a descriptor equation in the Value column using valid math and molecular operators.

For example, to create the equation for a descriptor that counts halogens in a structure, enter:

ecount(col "Structure", "Cl") + ecount(col "Structure", "Br")

This descriptor counts and reports the total number of chlorine and bromine atoms in a structure.

4.   In the Description column, enter a short description of the descriptor. For example, enter:

	Number of halogen atoms

for the description of the descriptor created in Step 3.

5.   In the 3D column, enter 0 if your descriptor is not a 3D descriptor. Enter 1 if the descriptor is 3D.

6.   In the Default column, enter 1 if you want the descriptor to be part of the default set. Enter 0 if the descriptor is not to be a default descriptor (Identifying default descriptors on page 168).

7.   In the Format column, enter the format for descriptor values to be displayed in the study table. The choices are float, integer, or scientific.

8.   In the Decimal column, enter the number of decimal places to be displayed in a descriptor value. If you entered integer in the Format column, enter 0.

9.   In the Units column, enter the units (for example, kcal/mol) to be applied to the descriptor value. If no units are to be applied, leave the cell blank.

10.   If the descriptor can be modified from a Cerius2 control panel, enter the name of the control panel in the Panel column. Otherwise, leave this column blank.

11.   To name the descriptor, click the first column in the row, then click the Prop (properties) tool. The Table Properties control panel appears. Select Row from the Properties popup.

12.   Enter a name (for example, Halogens) in the Row Name entry box.

13.   Click APPLY TO. The row name is entered in the first column of the selected row. QSAR+ sorts the descriptor list as it performs calculations, so the position of a descriptor in the list may change.

14.   Save the database containing the new descriptor. You can save the descriptor to the current database, to another existing database, or to a new database. For more information, Saving a descriptor database on page 171.

Note

When you finish creating a descriptor, you can check to see that it is correctly entered by adding it to the study table and inspecting the generated data (see Adding descriptors to the study table on page 154).

Modifying descriptors

ecount(col "Structure", "Cl") + ecount(col "Structure", "Br") + ecount(col "Structure", "F")

Save the database to activate the edited descriptor (see Saving a descriptor database on page 171).

When you finish modifying a descriptor, you can check to see that the modifications are correct by adding it to the study table and inspecting the generated data (see Adding descriptors to the study table on page 154).

Controlling the descriptor display format

To change the descriptor display format, edit the entry displayed in the Format column of the descriptor database table.

Creating new descriptor categories

Creating a new descriptor family

You can create new categories of descriptors by placing new entries in the Family column. For example, if investigator Jones wants to place all saved equations in a category named Jones-QSARs, Jones simply enters this designation in the Family column for the rows containing QSARs and saves the modified table. The value Jones-QSARs now appears as a choice in the family popup on the Descriptors control panel.

Saving a descriptor database

If you want to save the database that is displayed to the current database file, go to the study table or the QSAR card, select Descriptors/Databases to open the Descriptor Database control panel. Click the SAVE DATABASE pushbutton. The Save Database control panel appears, which lets you choose a name for your new or modified database.