Develop Pharmacophore Model Panel

Develop pharmacophore models from one or more ligands, a ligand-receptor complex, a receptor, a set of protein residues, or multiple existing hypotheses.

To open this panel, browse to click the Tasks button and browse to Ligand-Based Virtual Screening → Develop Pharmacophore Hypothesis.

  • Using
  • Features
  • Additional Resources

Using the Develop Pharmacophore Model Panel

This panel provides tools for developing pharmacophore hypotheses from a variety of sources and with a variety of methods. Each source and the methods used to develop hypotheses from it are described below.

Multiple ligands

When you have multiple ligands that are known to bind to a target, either strongly or weakly, you can develop a set of hypotheses that are optimized to represent the pharmacophore information derived from all the ligands. In this approach, the pharmacophore features on each ligand are identified, and a set of common features are sought that satisfy criteria on their positions and directions (if they are vector features), to form pharmacophore hypotheses. The hypotheses can be scored on their geometric alignment, their ability to retrieve actives from a set of decoys, and can be penalized for matches to inactive ligands. If multiple binding modes are a possibility, the hypotheses can be required to match fewer than the total number of active ligands, and the Detect Binding Modes Panel can be used to help determine which hypotheses match the possible binding modes. Excluded volumes can be added as a shell around the set of active ligands, or placed where inactive ligands have atoms where active ligands do not.

Single ligand

For a single ligand, you can build a hypothesis by selecting features on the ligand in the Workspace. You can also add an excluded volume shell around the ligand.

Receptor-ligand complex

If you have a receptor as well as a ligand, you can generate a hypothesis based on complementarity of receptor and ligand features, either automatically or manually. The automatic method involves using Glide XP scoring terms to determine which features contribute the most to binding (the e-pharmacophore model). or build a hypothesis by manually selecting features on the Workspace ligand, guided by interactions with the receptor. The manual method involves choosing features from a set that have complementary receptor features. For both methods, you can add excluded volumes that are based on the occupation of space by receptor atoms.

Receptor cavity

If you have a receptor but no ligand, you can build a pharmacophore model based on Glide XP scoring terms for fragments docked to the Workspace receptor (e-pharmacophore model). Like the receptor-complex method, the features are chosen to maximize the binding, but instead of a single ligand, fragments with the features are docked, and common features are chosen that satisfy criteria on their positions and directions. Clustering can be performed to reduce the number of fragments. You can also build a pharmacophore model manually based on inverse features from the receptor binding site.

Protein residues

If you are interested in protein-protein binding (such as for antibodies and antigens) you can build a pharmacophore model for a protein and use it in screening for binding to another protein. This method is essentially the same as for a single small-molecule ligand, for which you choose the desired features. As proteins can be large, you can restrict the choice of residues from which the features are selected to the residues of interest: the ones that bind to the other protein.

Merged hypotheses

Multiple hypotheses can be merged into a single hypothesis by picking features from the set of all features in the hypotheses. The hypotheses must be prealigned—for example, they may be derived from different ligands in the same binding site. This allows you to develop multiple models and choose common features manually.

The pharmacophore hypotheses are added to the project as entries. From there you can edit a hypothesis directly. Placing it in the Workspace and right-clicking on a feature opens the Edit Feature X Dialog Box. You can then reposition the feature, change the feature type, change the matching tolerance, decide which features it is allowed to and not allowed to match, how it is used in screening, or delete the feature. Clicking the H button for the hypothesis in the Entry List gives access to a range of actions you can perform on the hypothesis, including validation, adding to a screen, and managing excluded volumes. You can make measurements of distances, angles, and dihedrals between pharmacophore features, just like measurements involving atoms. You can show or hide the pharmacophore feature labels with the Annotations toggle on the Workspace Configuration Toolbar.

For information on running a workflow from the command line, see phase_find_common Command Help.

Develop Pharmacophore Model Panel Features

The features shown in this panel depend on the choice made from the Create pharmacophore model using option menu, as the options depend on what the pharmacophore model is based on. The options for each choice are described in their own section.

Create pharmacophore model using option menu

Choose the structures or hypotheses that you want to base the pharmacophore model on. The panel is updated with the features needed to make settings for the choice you made.

  • Multiple ligands—Develop a set of hypotheses from the set of ligands that is selected in the Project Table.

  • Single ligand—Build a hypothesis by selecting features on a single ligand in the Workspace. This is the default.

  • Receptor-ligand complex—Generate a hypothesis based on Glide XP scoring terms for the Workspace ligand docked to the Workspace receptor (e-pharmacophore model), or build a hypothesis by manually selecting features on the Workspace ligand, guided by interactions with the receptor.

  • Receptor cavity—Generate a model based on Glide XP scoring terms for fragments docked to the Workspace receptor (e-pharmacophore model), or on inverse features from the receptor cavity.

  • Protein residues—Build a model from the protein in the Workspace by selecting residues with the desired features.

  • Merged hypotheses—Build a consensus model from the hypotheses in the Workspace by picking features from different hypotheses.

Common features

The features that are common to several workflows are described below. The descriptions include any variation in features between workflows.

Feature selection tools

These tools allow you to select pharmacophore features using a grid or a table of checkboxes, or pick them in the Workspace. You can also add features by placing them in the Workspace.

Adding more than about 8 features to a pharmacophore model is not generally useful, and can increase screening resource requirements substantially.

Choose features checkboxes

Select the checkboxes for the features you want to include in the hypothesis. The features are marked in the Workspace. The checkboxes are presented in a grid if they come from a single structure. If the features can be selected from multiple structures or hypothesis, they are presented in a table that identifies the structure or hypothesis and the feature, with one feature per row.

Pick feature in Workspace option

Select this option and pick a feature in the Workspace to add it to the hypothesis.

Add feature in Workspace option and menu

Select this option to add a pharmacophore feature by placing it in the Workspace. Choose the feature type from the option menu and click in the Workspace to place the feature. To reposition the feature, right-click on it, and use the Edit Feature X Dialog Box to change its position, either by dragging or setting the coordinates.

Hypothesis Settings button

This button is present in all workflows except the merged hypotheses workflow. It opens the Hypothesis Settings Dialog Box, where you can make settings for the hypothesis features, scoring, and excluded volumes. The tabs and controls in this panel depend on the workflow.

Hypothesis name text box

Specify the name of the hypothesis to return. This feature is present when a single hypothesis is created. The name is also used for the job name when the single hypothesis is generated from a job.

Job name text box

Specify the name of the job that is run to generate hypotheses. This feature is present when multiple hypotheses are returned.

Create button

Create the hypothesis and add it to the Project Table.

Run button

Run a job to generate the hypothesis or hypotheses. For receptor-based hypotheses, the e-pharmacophore workflow is used to generate hypotheses. For ligand-based hypotheses, the job involves generating conformers, aligning ligands, finding common pharmacophores, scoring and selecting hypotheses. The hypotheses are added to the Project Table when the job has finished.

Settings button

This button has a menu with two items that are common to all workflows: to reset the settings for the current task (workflow) to their defaults, and to reset the settings in the entire panel to their defaults. For workflows that run jobs, the button opens the Job Settings Dialog Box, and the menu has other items that are described in Job Toolbar.

Job status button

This button indicates the status of a job in the workflows that run jobs. See Job Toolbar for details.

Multiple ligands workflow

The unique features for this workflow are described below. The links for the common features take you to the description in the Common features section.

Actives / Inactives Split section

This section provides information about and tools for the division of the ligands into actives and inactives. The active ligands are used to develop the pharmacophore model; the inactive ligands are used for scoring and creating excluded volumes. In the resulting pharmacophore model, the inactives are divided into two groups: those that match all sites (Inactives Full Match), and those that match some but not all of the sites (Inactives Partial Match).

Report text

Reports the number of ligands assigned to the active set and the inactive set.

Define button

Specify the ligands to include in the active set and the inactive set. Opens the Define Actives and Inactives Dialog Box.

Pharmacophore method options

Select a method for creating a pharmacophore model:

  • Find best alignment and common features—align the ligands so that common features coincide. The default scoring function for the hypothesis, with this choice, is the Phase Hypo Score function.
  • Use prealigned ligands—the ligands are already aligned so that the common features coincide. The default scoring function for the hypothesis, with this choice, is the custom function.
Generate conformers option

Select this option to generate conformers for the ligands. If the entries you chose include contiguous sets of conformers, these are detected automatically and used as the conformers for the model development.

Options button

Set options for the generation of the conformers. Opens a dialog box in which you can specify the target number of conformers to generate, and choose to minimize the conformers.

Single ligand workflow

The features for this workflow are all described in the Common features section.

Receptor-ligand complex workflow

This workflow has two methods for generating pharmacophore models: automatic and manual, with different options. The automatic workflow does not have the options for showing features or feature selection.

The unique features for this workflow are described below. The links for the common features take you to the description in the Common features section.

Method options

Choose a method for generating a pharmacophore model from the complex.

  • Auto (E-pharmacophore)—Choose the pharmacophore features based on Glide XP scoring terms (e-pharmacophore method). The best-scoring features are included in the pharmacophore model. The number of features can be defined in the Hypothesis Settings Dialog Box.
  • Manual—Choose the pharmacophore features from a list of features on the ligand that have complementary features on the receptor.
Choose ligand list

Choose the ligand that the pharmacophore model is based on from this list, which shows all molecules in the Workspace that meet the criteria for a ligand. These criteria are defined in the Ligand Detection Preferences of the Preferences Panel.

Show features slider

Change the number of features that are shown in the Workspace. If the total number of features is large, you can use this slider to restrict the number of features shown. Ligand features or inverse features are shown when they have complementary features in the receptor that are within prescribed distance and angle cutoffs. The number of features is controlled through the cutoffs. Only available with the manual method.

Show inverse features from receptor option

Show "inverse" features in the Workspace and the table instead of the ligand features. The inverse features are derived from the complementary receptor features, placed at a location where a ligand feature would have a favorable interaction with the receptor.

Receptor cavity workflow

In this workflow, the receptor binding site must be defined in terms of residues or coordinates, as there is no ligand. An e-pharmacophore job is run to dock fragments to the receptor, and the best binding fragments are used to define the pharmacophore features.

The unique features for this workflow are described below. The links for the common features take you to the description in the Common features section. The presence of some of the features depends on the method chosen.

Method options

Choose a method for generating a pharmacophore model from the receptor cavity.

  • Auto (E-pharmacophore)—Choose the pharmacophore features based on Glide XP scoring terms from fragment docking (e-pharmacophore method). The best-scoring features are included in the pharmacophore model. The number of features can be defined in the Hypothesis Settings Dialog Box.
  • Manual—Choose the pharmacophore features from a list of inverse features from the receptor cavity. The inverse features are derived from the complementary receptor features, placed at a location where a ligand feature would have a favorable interaction with the receptor, possibly with multiple receptor features.
Define the receptor binding site using options

Choose an option for defining the center of the binding site. The grid box for the Glide docking run used in this workflow is centered at this location.

Centroid of residues option and Specify Residues button

Center the Glide docking grids at the centroid of a set of residues that you select. Click the Specify Residues button to open a dialog box in which you can choose the residues. This dialog box is the same as the Glide Active Site Residues Dialog Box.

X,Y,Z coordinates option and text boxes

Enter the coordinates of the grid center in these text boxes. If using the manual method, click Apply to apply the grid center definition and generate the inverse features.

Show features slider

Change the number of features that are shown in the Workspace. If the total number of features is large, you can use this slider to restrict the number of features shown. The number of features is controlled through cutoffs on distances and angles that define the likely ligand feature locations and orientations. Only available with the manual method.

Protein residues workflow

In this workflow, the pharmacophore model is constructed from protein residues in a protein-protein system, where one protein functions as the ligand. This workflow is useful for screening proteins for docking to other proteins (such as antigen-antibody docking).

The unique features for this workflow are described below. The links for the common features take you to the description in the Common features section.

Specify Residues button

Click this button to select the residues that contain the pharmacophore features you are interested in. The dialog box that opens is the same as the Glide Active Site Residues Dialog Box. Once the residues are selected, the Choose features area is populated with toggle boxes for the features on these residues.

Merged hypotheses workflow

In this workflow, the pharmacophore features come from existing hypotheses rather than from a set of ligands. It is assumed that the hypotheses are aligned in the same way to the reference. The task is to choose features that are in approximately the same location in all the hypotheses.

The tools available are all common. The links take you to the description in the Common features section.

Tutorials

Related Topics