View Hypothesis Workbench

Contents:

Illustration of View Hypothesis workbench
Introduction to the View Hypothesis workbench
Tasks you can do in the workbench

Menu descriptions:

Illustration of View Hypothesis workbench

Introduction to the View Hypothesis workbench

You can use the View Hypothesis workbench to create, view, modify, and evaluate hypotheses. A hypothesis is a collection of 3D, 2D, and 1D (property) specifications that describe the distinguishing physical and chemical characteristics of a class of molecules. You can use a hypothesis to describe molecule characteristics in very general or very precise terms. Then you can use the hypothesis to search a database.

For example, a hypothesis might represent the generic class of barbiturates or might represent a pharmacophore for a beta 2 antagonist such as a phenyl ring, a primary alcohol, and a secondary nitrogen, all separated by distance constraints.

For a more detailed discussion of what hypotheses are and what you can use them for, please see Introduction to hypotheses. For illustrated examples of hypotheses, see Examples of Hypotheses.

What is a hypothesis made of?

A hypothesis may contain any combination of features (e.g., chemical functions and topological fragment specifications), constraints involving features, and property (1D) specifications:

Functions

A hypothesis function describes a specific chemical property or behavior, such as a hydrogen-bond donor or a positive charge. Catalyst has eight predefined chemical functions that are available from the Feature Dictionary and that are used in speeding up database searching: hydrogen-bond acceptor, hydrogen-bond acceptor (lipid), hydrogen-bond donor, hydrophobic, negative charge, positive charge, negative ionizable, and positive ionizable. Automatic hypothesis generation uses these functions by default to define hypotheses. For more detailed information, please see Description of Catalyst's predefined chemical functions.

You can create customized chemical functions with the editing tools in the View Hypothesis workbench. You can modify an existing function by defining additional allowed topology or additional excluded specifications for a function using the Exclude/OR editor. Or you can define a feature as either a phenol group or a benzoic acid group. The Exclude/OR editor is described under The Exclude/Or QuickTool.

Illustrations of two predefined functions:

Fragments

A hypothesis fragment is a substructure made up of specifications for atoms and bonds. You can specify each atom in the fragment as a particular element or can specify that it may be any of several allowed elements. Similarly, each bond may be specified as a particular bond type or you can allow it to be any of several types. A topology hypothesis can be created by converting a molecule to a hypothesis or by selecting a fragment from the Feature Dictionary in the View Hypothesis workbench. Although a fragment hypothesis may look the same on the screen as a molecule, it has the properties of a hypothesis. See How Catalyst displays hypothesis fragments.

Constraints

Catalyst uses several types of constraints in hypotheses:

Geometric constraints. You can define constraints on location, distance, angle, and torsion in hypotheses. Each type of constraint can be used in a hypothesis to specify physical limitations on the characteristics of target molecules. Of this group, location constraints are the most important part of a hypothesis in defining the geometry of target molecules.
Geometric objects. To assist in defining geometric constraints, you can define a geometric centroid or other type of point, a vector defined by a head and a tail point, a hydrogen-bond donor or acceptor, or a plane defined by a least-squares fit to a set of points (with its normal vector). These objects can be used to specify constraints on target molecule characteristics.
Excluded-volume constraints. You can place an excluded-volume constraint in a hypothesis, which specifies a spherical volume in which no objects are allowed for geometrical comparisons or database searches.
Properties. You can assign 1D specifications in hypotheses; these are scalar values such as an allowable range for the molecular weight.

How Catalyst displays hypothesis fragments

In a hypothesis, each fragment looks like a collection of atoms and bonds using the same color conventions as for molecules. However, the objects in hypotheses that look like atoms are specifications for atoms rather than atoms. For example, an atom in a hypothesis might represent "any halogen", "any metal", or "one carbon or nitrogen or oxygen". When an atom in a hypothesis has several possible identities, its color is mauve. When a hypothesis feature represents a more complex specification than can be displayed easily, it is shown surrounded by a yellow mesh. You can see the specifications that are hidden by displaying the complex hypothesis in the Exclude/OR editor.

If an atom in a hypothesis looks the same as in a molecule, this means that it is a single-atom specification. That is, a red atom would specify an "oxygen only".

One common specification for an atom in a hypothesis is that it can be "any element except hydrogen". For example, if you add a secondary amine to the workspace, you see a structure consisting of one nitrogen bonded to one hydrogen and two brownish-mauve atoms:

Here, each brownish-mauve atom represents an atom that may be any element except hydrogen. A secondary amine must, by definition, be bonded to one hydrogen and to two non-hydrogen atoms.

The number of displayed hydrogens represents the minimum hydrogen count that is permitted. Therefore, if "any" hydrogen count (0-8) is specified, no hydrogens are displayed.

Tasks you can do in the workbench

Build hypotheses

Modify hypotheses

To merge multiple items into one feature or hypothesis
To delete components of a hypothesis
Set specifications for bonds and atoms
Specify constraints for a hypothesis
- To create and modify distance constraints
- To add location constraints to functions and atoms
- To add angle constraints
- Torsion-angle constraints
- Excluded-volume constraints
- Defining geometric objects (such as centroid, vector, hydrogen-bond donor or acceptor vectors, and best-fit plane)
Use the Exclude/OR QuickTool to add alternative features and exclusions to a hypothesis

Use hypotheses

Bring in an existing hypothesis and save or export a hypothesis
To score a hypothesis
To search a database or spreadsheet for compounds that match a hypothesis
Compare/Fit it with a compound or another hypothesis
Regress it to a training set of molecules, so that it can be used in estimating activity
Estimate the activity of a molecule

To build a hypothesis

If you have good information about the chemical and geometric characteristics of your desired hypothesis, then, as an alternative to automatically generating a hypothesis, the View Hypothesis workbench provides three ways for you to manually construct hypotheses:

Use a template molecule to place functions

Suppose you have a molecule that represents several chemical functions that you want to include in your hypothesis. You can use the molecule as a template to define the locations of chemical functionalities (rather than to define the specific topology of a hypothesis).

For example, you may have a compound that has three hydrophobic groups in the proper relative locations, and you want to build a hypothesis that has three hydrophobic groups that have the same spatial relationships. You can use the molecule as a template to build a hypothesis in the View Hypothesis workbench. Catalyst can identify the locations of the chemical functions (such as hydrophobic or positive ionizable groups) in the molecule and also allow you to choose the specific bond direction for vector functions such as hydrogen-bond acceptor or donor. Then you can add the desired functions at those locations and merge them to form a hypothesis with the exact geometry you want.

Convert a molecule to a hypothesis

You can bring a molecule containing the topological features you want into the View Hypothesis workbench and convert it into a hypothesis. You can thus build a fragment hypothesis consisting of any topology that has the characteristics you want. You can then modify atom and bond specifications using the atom and bond specification editors.

A fragment hypothesis is considered a single-feature hypothesis, no matter how complicated its topology. No chemical functions are inherently recognized in a fragment hypothesis, regardless of what atom specifications are included. Chemical functions must be added if you need them in your hypothesis. In general, you should not have both a fragment and a function in one hypothesis that represent the same feature.

Add features from the Feature Dictionary

You can pick functions and fragments from the Feature Dictionary (see Using the Feature Dictionary to build hypotheses) and place them in the workspace. This method allows you to construct a hypothesis from a set of predefined building blocks that include many types of chemical groups, such as phenyl rings or carboxylates, and also chemical functions, such as hydrogen-bond donors or positive charges. However, you cannot directly specify the 3D coordinates for features placed using this method, so it is not recommended for hypotheses in which 3D coordinates and location constraints are important. For this reason, the other two methods generally are preferred for manually constructing hypotheses. This method can be used successfully for features and constraints that can be described with 2D specifications, such as distance or angle constraints.

To finish building your hypothesis

After using any of the above three methods of constructing a hypothesis, you need to assign weights and location constraints to some or all the functions of the hypothesis, so you can use the hypothesis in compare/fit or database search operations. You will need to use the Regress Hypothesis menu item with a relevant training set of compounds to be able to estimate activities with your hypothesis.

More information on constructing hypotheses:

More information on refining hypotheses after the initial construction:

To merge multiple items into one feature or hypothesis
To delete components of a hypothesis
Set specifications for bonds and atoms
Set feature weight
Specify constraints for a hypothesis:
- To create and modify distance constraints
- To add location constraints to functions and atoms
- To add angle constraints
- Torsion-angle constraints
- Excluded-volume constraints
- Defining geometric objects (for defining constraints such as centroid, vector, projected hydrogen-bond donor or acceptor points, or best-fit plane)
Use the Exclude/OR QuickTool to add alternatives and exclusions to a function

Using the Feature Dictionary to build hypotheses

The Feature Dictionary in the View Hypothesis workbench provides predefined hypothesis features of two types:

Fragments (substructures)
Chemical functions

Either or both sets of features can be displayed in the Feature Dictionary list box by selecting the Functions Only, Fragments Only, or Fragments & Functions button. Functions and fragments can be selected from the list and placed in the workspace. You also can create your own feature definition in the View Hypothesis workbench and add it to the Dictionary for future use (see the description of the Add to dictionary menu item).

Fragments

Each fragment in the Feature Dictionary is a single-feature hypothesis representing specifications for a topological group of atoms and bonds.

Chemical Functions

Chemical functions in the Feature Dictionary are hypotheses specifying certain chemical behaviors. The available functions are:

HB ACCEPTOR: Hydrogen-bond acceptor (vector function)
HB ACCEPTOR lipid: Hydrogen-bond acceptor in a lipid environment (vector function)
HB DONOR: Hydrogen-bond donor (vector function)
HYDROPHOBIC: Hydrophobic group (point function)
HYDROPHOBIC aliphatic: Hydrophobic aliphatic group (point)
HYDROPHOBIC aromatic: Hydrophobic aromatic group (point)
NEG CHARGE: Negatively charged group (atom specification)
NEG IONIZABLE: Group that is, or can be, negatively charged -- for example, carboxyl (point function)
POS CHARGE: Positively charged group (atom specification)
POS IONIZABLE: Group that is, or can be, positively charged -- for example, nonaromatic amines (point function)
RING AROMATIC: Five- or six-membered aromatic ring (vector)

For more details, please see Description of Catalyst's predefined chemical functions.

How Catalyst displays hypothesis functions

Hypothesis functions are represented in different forms, depending upon whether they have the characteristics of atom specifications, point functions, or vector functions:

Atom specifications, such as positive and negative charge, are represented as solid mauve spheres with a yellow mesh around them. The yellow mesh indicates that the object has additional specifications that can be viewed in the Exclude/OR editor.
Point functions represent chemical behaviors that may not be located at an atom, such as hydrophobic and positive and negative ionizable functionality, and are displayed as light blue spheres.
Vector functions, such as hydrogen bond acceptors and donors, are represented as a mauve sphere at the location of the heavy atom, a light blue arrow representing the direction of the hydrogen bond, and a light blue sphere representing the projected point receptor site. The heavy atoms have a yellow mesh around the atom, which indicates that the object has additional specifications that can be viewed in the Exclude/OR editor.

Predefined Catalyst hypothesis functions

Please see Description of Catalyst's predefined chemical functions for detailed information about each function.

To add items from the Dictionary

To use the Feature Dictionary to add items to a hypothesis:

Open the View Hypothesis workbench, if it is not already open, by double-clicking the View Hypothesis button in the tool bar.

The Feature Dictionary windown appears at the top right of the screen. Bring the Feature Dictionary to the front by clicking the edge of any of its borders or by selecting it from the Windows menu.

You can use this dictionary to add items to your hypothesise.
The dictionary allows you to select a fragment or a function. You can choose whether only fragments, only functions, or both are displayed in the scrollable menu. For a discussion of functions and fragments, see What is a hypothesis made of?.
To display only fragments, choose Fragments Only.
To display only functions, choose Functions Only.
To display both, choose Fragments & Functions.
Select the function or fragment to add by clicking its name in the list box. The name is highlighted.
You can quickly select a fragment or function whose name contains certain letters, as explained in How to search the Feature Dictionary.
Click anywhere in the 3D workbench to place the selected fragment or function in it.
The selected function or fragment appears in the 3D workbench. You can move it, zoom it, rotate it, and save it to the shelf, using the same techniques as in any other workbench. You can merge it with other features to create more complex hypotheses.

You also can create your own feature definition in the View Hypothesis workbench and then add it to the Dictionary for future use. See the description of the Add to Dictionary menu item.

How to search the Feature Dictionary

You can search to find a particular entry in the Feature Dictionary quickly. The case of the letters in your search string must match that of the target list-box item (functions are listed in uppercase and fragments have initial caps). The search first tries to match the exact string entered. If that fails, the search tries to find the string with a trailing wildcard. If that is unsuccessful, it searches again, with both leading and trailing wildcards.

To search for an entry in the Feature Dictionary:

Select the appropriate Fragments or Functions form of listing for the feature you want.
In the Search String box, enter the string to search for.
For example, you can enter Carbox or box to match Carboxyl in the list box, C to scroll to the first name that starts with C, or a to scroll to the first name that contains a lowercase a.
The search always starts at the top of the list box, so you always get the same fragment or function with a given search string.
Press <Enter> on your keyboard or select the Search button in the Feature Dictionary.
The list box scrolls to the first name that contains the exact string you entered. The full name becomes selected and highlighted.

Editing techniques

Deselecting selections

To deselect selected atoms, bonds, functions, and fragments in the View Hypothesis workbench, click a selected item in the workspace, use the region-select technique to draw a small rectangle in an empty area of the 3D workspace, or click the Deselect button in the toolbox.

Undoing mistakes

If you accidentally add an unwanted function or fragment to the workspace, leave it selected and click the Erase tool in the toolbox to delete it.

Modifying your hypothesis

You can modify a hypothesis by modifying atom and bond specifications and by setting and modifying distance, angle, torsion, and location constraints.

You can add more items to the workspace by selecting and adding them from the dictionary or by converting molecules into hypotheses. Each item comes into the workspace as a separate hypothesis, and you must merge them to join them into one hypothesis.

More information on modifying hypotheses:

To merge multiple items into one feature or hypothesis
Set specifications for bonds and atoms
To create and modify distance constraints
To add location constraints to functions and atoms
To delete components of a hypothesis

To convert molecules to hypotheses

You can bring a molecule into the View Hypothesis workbench and convert it to a single-feature topological hypothesis. This allows you to create hypotheses containing any chemical structures you like. You can build hypotheses made entirely from items in the Feature Dictionary or entirely from molecules converted to hypotheses or from any mixture of the two.

To bring a molecule into the shelf of the View Hypothesis workbench, you can:

Import a molecule from a file.
Drag a molecule from a lab or a workbench and drop it onto the workbench shelf.
Use the View Compound QuickTool to create a molecule (see How to use the View Compound QuickTool).
Use the Data/Find menu item to find molecules in a database and then copy the results to the shelf. This method is useful when you want to build a hypothesis that incorporates characteristics of a compound that exists in a database (see The Find menu item).

To convert a molecule into a single-feature hypothesis:

Bring the desired molecule into the workspace of the View Hypothesis workbench, using one of the above methods.
If you want to convert more than one molecule at a time, bring the other molecules into the workspace and select all molecules you want to convert.
If there is only one molecule in the workspace, you do not need to select it.
Select the Convert Molecule to Hypothesis button in the toolbox.

The molecule becomes a single-feature hypothesis. Some visual changes that occur after the conversion include:
- The rings indicating aromatization disappear.
- If you display atom labels, the atom names appear as values with { } brackets around them.
You can now modify the hypothesis. You can set and modify atom and bond specifications and place constraints and geometric objects. You can delete atoms, bonds, and constraints by selecting the items to be deleted and then selecting the Erase tool in the toolbox. You can add a weight to the hypothesis. Unlike a hypothesis with several features or functions, only one weight can be assigned to this hypothesis, because the converted substructure is a single feature.
You can add more items to the workspace by selecting and adding them from the dictionary or by converting molecules into hypotheses. Each item comes into the workspace as a separate hypothesis, and you must merge them to join them into one hypothesis.

More information:

To merge multiple items into one feature or hypothesis
Set specifications for bonds and atoms
To create and modify distance constraints
To add location constraints to functions and atoms
To delete components of a hypothesis
Defining geometric objects to add to a hypothesis

To delete components of a hypothesis

To delete atom or bond specifications and constraints in hypotheses, select them in the workspace and then select the Erase tool in the toolbox.

To merge multiple items into one feature or hypothesis

When you add a new function or fragment to the workspace, it is a separate hypothesis, not connected to anything else. You can merge several different features into one hypothesis having separate features or into one hypothesis having just one feature. To maintain the separateness of a feature, set a weight and/or a location constraint on the feature. (If separate features do not have weights or location constraints, they are merged into one feature in the merged hypothesis.)

To merge separate features or hypotheses into one hypothesis:

Select all items to be merged.
From the toolbox select the Merge Features tool.

Merging specifies that all features must be present to satisfy the hypothesis. However, merging does not create any specific distance relationships between the merged features unless they have location constraints. If you want to specify spatial relationships between features you also can specify distance, angle or torsion constraints (as discussed under To create and modify distance constraints) between the features.

Or you can set bond specifications between atom features (see the description of the Set Bond between Features tool). If you set spatial relationships between features by means of constraints or bond specifications, merging the features is not necessary. The features are joined into one hypothesis by the distance, angle, or torsion constraints or the bond specifications.

Specifying constraints

To create and modify distance constraints

You can specify distance constraints between appropriate types of items. For example, you can specify that a particular atom must be 5-15 Å from another specific atom, from a hydrogen-bond acceptor, or from a defined plane (geometric object). You can add distance constraints to a centroid (center of geometry) of a ring that you have defined as a geometric object.

When you specify a distance constraint between two items in separate hypotheses, the hypotheses automatically become merged into one hypothesis.

More information:

To add distance constraints
To change constraint values and names
To display constraint values and names

To add distance constraints

To specify the distance range between two components in a hypothesis:

Use the extend-select technique to select two components in the workspace.
Select the Constraints/Define Constraint/Distance menu item.
The distance constraint is indicated as a white line with yellow highlighting, drawn between the two selected items, as shown below. The Constraint_Tolerance control panel also appears:
The Name box shows the default name of the constraint. Change it if you want, by deleting the default and entering the new name.
The default value for Constraint Minimum (minimum distance) is the current distance between the selected features minus one angstrom, and the default Constraint Maximum is the distance plus one angstrom. You may change these values if you want.
Select Apply to implement the changes while keeping the control panel on the screen or select OK to implement the changes and close the control panel.

To display constraint values and names

To display the values of constraints, use the View/Constraint Tolerances menu item . The values are displayed, in the units of measurement selected in the Global Preferences control panel (Preferences menu in the Stockroom or lab).

To display constraint names, use the View/Constraint Names menu item.

These menu items are toggles. You switch the display of the constraint values and names on and off each time you select the menu items.

To change constraint values and names

To change the values for an existing constraint or to change the name of a constraint, use the Set Constraint Tolerance tool in the toolbox.

Select the constraint in the View Hypothesis workbench.
Go to the toolbox and click the Set Constraint Tolerance button.
A control panel appears. It is the same control panel that appeared when you first created the constraint.
Change and values for the constraint that you want to.
Change the constraint name if you want to.
Select the OK button to apply the changes and to close the control panel or select the Apply button to apply the changes while keeping the control panel on the screen.

To change the display style of hypothesis objects

You can change the appearance of hypotheses and molecules in the 3D workspace in a variety of ways:

Display parameters that you can change How to change them

Constraint display Use the Style/3D Constraints menu item. The choices are Invisible, Mesh, Dots, Transparent, Solid.

Geometric object display Use the 3D Geometric Objects menu item. The choices are Invisible, Mesh, Dots, Transparent, Solid.

Constraint names Use the ViewConstraint Names menu item.

Geometric object names Use the View/Geometric Object Names menu item.

Constraint tolerances Use the View/Constraint Tolerances menu item.

Fragment display Choose a style from the Style menu. The choices are Wireframe, Mesh, Dots, Tube, Ball and Tube, Sphere and Rod, Spacefilling.

Molecule display Choose a style from the Style menu. The choices are Wireframe, Mesh, Dots, Tube, Ball and Tube, Sphere and Rod, Spacefilling.

Atom labels Use the View/Atom Labels menu item.

Function names Use the View/Atom Labels menu item.

Hypothesis names Use the View/Compound Names menu item.

Compound names Use the View/Compound Names menu item.

Display parameters that you can change	How to change them
Constraint display	Use the Style/3D Constraints menu item. The choices are Invisible, Mesh, Dots, Transparent, Solid.
Geometric object display	Use the 3D Geometric Objects menu item. The choices are Invisible, Mesh, Dots, Transparent, Solid.
Constraint names	Use the ViewConstraint Names menu item.
Geometric object names	Use the View/Geometric Object Names menu item.
Constraint tolerances	Use the View/Constraint Tolerances menu item.
Fragment display	Choose a style from the Style menu. The choices are Wireframe, Mesh, Dots, Tube, Ball and Tube, Sphere and Rod, Spacefilling.
Molecule display	Choose a style from the Style menu. The choices are Wireframe, Mesh, Dots, Tube, Ball and Tube, Sphere and Rod, Spacefilling.
Atom labels	Use the View/Atom Labels menu item.
Function names	Use the View/Atom Labels menu item.
Hypothesis names	Use the View/Compound Names menu item.
Compound names	Use the View/Compound Names menu item.

[Top] [Index]