R-Group Analysis — Importance Analysis Panel

The importance analysis panel displays information on the range of property values for each attachment position and the values of the properties for each of the R groups at a selected position.

To open this panel, click Importance Analysis in the R-Group Viewer Panel, or choose Show R Groups from the shortcut menu in the histogram display of the Heat Map Panel.

  • Overview
  • Features
  • Additional Resources

Overview of Importance Analysis

Importance analysis addresses the question of how sensitive the property of interest is to R-group variation at that position. A position is more important than another if varying the R group at that position leads to greater property differences than are observed when the R group at the other position is varied. The importance value of a position is the range of the property over the R groups at that position, averaged over all structures containing a given R group at that position. As an example of its use, a position in which the range of the property does not vary much as the R group is changed may not be very important for improving the potency of a lead compound, whereas a position in which the property varies a great deal may be considered much more important.

Ideally, we would like to interpret the importance measure as characteristic of a position. For example, positions that make intimate contact with a receptor might be expected to exhibit greater variation of a binding property when its R groups are altered than one which protrudes into the solvent. However, there are several caveats to this interpretation. Most SAR data sets have very different numbers of structures at the various positions. Even if all else were equal, we would expect positions with more R groups to exhibit greater property variation and therefore to appear more important. Even if all positions had the same number of R groups, we would expect positions whose R groups exhibit greater chemical variability to exhibit a greater range of property values upon variation and hence a greater importance. For example, the chemists designing the SAR study might have had reason to believe that it is important to place hydrogen-bond donors at a given position. They might then sample only such R groups there, and this could in some cases lead to minimal property variation and low importance. These possibilities should be kept in mind when interpreting the importance measure.

Importance Analysis Panel Features

 

Displayed property section

In this section you choose a property for the importance analysis and the function of the property that is used to define the values for each R group.

Average, minimum, maximum option menu

Choose the function of the property that is used to calculate the values to display for each R group. The available functions include maximum, minimum, median, average, and standard deviation. For each table cell, the function is applied to the values of the property for all molecules that contain the R group represented in that cell (which implies looping over all other positions). The R group table is updated when you choose an item from this menu.

Property option menu

Choose the property that is used to calculate the importance and the values displayed in the table from this option menu. The histogram and the R group table are updated when you choose an item from this menu, showing the most important position for that property.

Cores display

This section displays the core structures. If you pause the pointer in a cell, an enlarged image of the core is displayed.

R-Group position importance display

In this section, the relative importance of the positions for the chosen property is displayed as a histogram, which you can use to choose the position for which R groups are displayed in the table below. The selected position is marked with an asterisk on the X axis.

The relative importance of a position for a particular property is defined as the range of values of the chosen property among the R groups at that position, where the value is the average over all other positions. The importance values for the positions are scaled so that the largest value is 1.

R-Group table

This table displays the R groups at the position chosen in the histogram, and the value of the selected function of the property, evaluated over all R groups at the other positions. If you click in a cell, the Structure panel opens, and allows you to display the structures represented in that cell, with the value of the property that is chosen for display. You can sort the property values in ascending or descending order, by clicking on the column heading.

Legend and Settings button

Change the representation of the values displayed, whether colors or values are displayed, and the range of the color ramp. Opens the Display Settings Dialog Box.

Cell display option menu

Choose the representation of the property value that is displayed in the table cells.

  • Values and colors—show the property value and the color.
  • Cartwheels and colors—show a cartwheel colored by unique property values (see below), with the average value for the cell background.
  • Cartwheels only—show a "cartwheel" in which the sectors of the circle represent unique values of the property, and are colored according to the property value.
  • Values only—show the property value only
  • Colors only—show only the color corresponding to the property value

When values are displayed, the number of structures for which the property is evaluated and the total number of structures for that cell are displayed in parentheses below the value.

Cell size slider and text box

This slider controls the size of the cells in the heat map. The size is displayed in the text box, which you can edit directly to change the cell size, or use the arrow buttons.

Export buttons

These buttons allow you to export the histogram and the table data.

  • Export Graph—Export an image of the histogram, in PNG, TIFF, or JPEG format.
  • Export Image—Export an image of the R-group table, in PNG, TIFF, or JPEG format.
  • Export CSV—Export the R-group table data to a CSV file with a header, with the structure represented as a SMILES string. The SMILES string includes stereochemistry, and marks the attachment points with [Xe] and [Kr] (as the nominal capping group).

Related Topics