Tools Menu

Select A Command:

Score Hypothesis

Generate Hypothesis...

Regress Hypothesis

Sort by Property...

Show Selected Compounds/Mappings

Show Conformational Model...

Compare/Fit...

• two compounds - performs an RMS geometric comparison of selected atoms in the two compounds, including Fast Fit or Best Fit on the conformers of one or both compounds (in View Compound workbench)

• a compound and a hypothesis - for a hypothesis with feature weights, checks that geometric constraints are met and then performs a weighted geometric fit of matching features. For a hypothesis without feature weights, performs an RMS geometric comparison of selected features in the two objects

• two hypotheses - performs an RMS geometric comparison of matching features

Score Hypothesis

Score Hypothesis requires as input a set of lead molecules that all have conformational models, a related hypothesis, and a saved spreadsheet containing a tested activity value and the uncertainty in the activity for each compound.

To use the Score Hypothesis command:

1. Make sure that a conformational model has been generated for all compounds to be used in Score.

2. Drag the lead set compounds into the tabular report in the Generate Hypothesis workbench. Or import a spreadsheet that references the compounds to be used.

3. Prepare and bring the hypothesis to be scored into the workbench shelf.

4. Type in the activity for each compound in the lead set, if not already entered. Modify the default uncertainty values, if necessary. For more details about preparing a spreadsheet, see "Introduction to the Tabular Report/Spreadsheet."

5. After the data are all entered, save the spreadsheet with the Save Spreadsheet command on the Data menu.

6. Select the hypothesis and the spreadsheet on the shelf.

7. Select Score Hypothesis on the Tools menu. The program produces:

   An estimated activity for each compound is displayed the Est column, and the ratio between the estimated and tested activity in the Error column.

   The Status Area reports the Score value, indicating how well the hypothesis fits the compounds. The Score value is a root-mean squared error in the estimated activities, so a low score (less than 1.0) represents a good fit.

   See "To Score a Hypothesis" for more details.

The functions used by default in hypothesis generation are the same as those in the Feature Dictionary available for constructing hypotheses manually:

• Hydrogen bond acceptor
• Hydrogen bond acceptor (lipid)
• Hydrogen bond donor
• Hydrophobic
• Negative charge
• Negative ionizable
• Positive charge
• Positive ionizable

Initially, you must bring all of your lead set compounds into Catalyst, generate a conformational model for each one, then prepare an input spreadsheet using the tabular report in the Generate Hypothesis workbench. You must enter their tested activity values in the report, and modify the default uncertainty value, if desired, and save the spreadsheet. Select Generate Hypothesis and the dialog box is displayed.

After choosing the set of functions you want to consider and the machine on which to run, you set up a background process to perform the generation. When completed after a number of hours, bring your results into Catalyst for review.

For more details about the procedure to set up and generate a hypothesis, see a list of hypothesis generation tasks.

Regress Hypothesis...

The resulting hypothesis can be used to estimate activities of molecules that have behavior similar to the training set used in Regress.

To use Regress Hypothesis,

1. Drag the training set compounds to be used to generate the regression equation into the tabular report in the Generate Hypothesis workbench. Make sure that all compounds have a conformational model.

2. In the tabular report, type in the tested activity for each compound in the Activ column (if a value is not already available). Modify any activity uncertainty value in the Uncert column, if you do not want to use the default value (3.0). The uncertainty represents a ratio range for the actual activity. For example, for a compound with an activity of 30, an uncertainty of 3.0 allows an activity range of 10 to 90.

   Note that any changed values are displayed on a grayed background to indicate that the changes have not been committed to the StockroomDB.

3. Save the spreadsheet to the shelf with the Save Spreadsheet command from the Data menu.

4. Bring the hypothesis for which you want to generate a regression equation into the shelf of the workbench.

5. Select the spreadsheet and the hypothesis on the shelf.

6. Select the Regress Hypothesis command from the Tools menu.
   When the program is finished, a new local hypothesis is displayed on the shelf, and the following results are displayed.

In the Error column, the ratio of tested/estimated activity is displayed with a negative sign if the tested activity is higher than estimated. The Error column displays the ratio of estimated/tested activity with a positive sign if the estimated activity is higher than the tested activity. For example, if Score displays an Error of -1.57 for a compound, the tested activity is 57% lower than the activity estimated with the hypothesis.

An Xgraph is displayed showing the regression relationship of estimated activity with tested activity for the lead set.

Sort by Property...

1. Select the Sort by Property command from the Tools menu.

   The Sort By Property dialog box is displayed, with a list of properties that can be used to sort the tabular report entries. Use the scroll bar at the right of the list to see properties that are not visible.

2. Click on the P symbol next to the name of the property to be used to sort the list.

   The border of the selected property is highlighted.

   In the Description text box, the selected property is described.

   In the Dictionary text box the property dictionary is identified-either Stockroom, Corporate, or Biocad.

   In the Type text box, the data type of the selected property is identified-either real, integer, string, or date.

3. Select one of the Direction toggle buttons to choose the method to sort:

   1, 2, 3, 4... sorts entries in normal ascending order

   9, 8, 7, 6... sorts entries in descending order

4. Select the Sort button to start sorting and leave the dialog box displayed.

Select the OK button to start sorting and dismiss the dialog box.

The entries in the tabular report are reordered according to the Property and Direction selected.

Show Selected Compounds/Mappings

1. Generate conformers for any compounds you want to evaluate that do not have them.

2. Bring the hypothesis and the related compounds into the Generate Hypothesis workbench. The molecules can be from the training set used to generate the hypothesis, or they may be related compounds with similar binding behavior that were not included in the training set.

3. Drag the hypothesis into the Current Hypothesis text field, or select the hypothesis on the shelf and then select the Set Current Hypothesis command from the Edit menu.

4. Extend select and drag the molecules into the report area. (If you want to, you can rearrange their order by using the Sort by Property command from the Tools menu.)

   Each compound is assigned a unique color (for displaying in the 3D workspace superimposed with other compounds) shown in the Color column. If you want to change any color, click on the color square and the Compound Overlay Color dialog box is displayed. Select the color you want from the color wheel and then select OK. The display color for the compound is changed.

5. To do a quick compare/fit of a compound to a hypothesis, click on the row number of the compound in the spreadsheet and then select the Show Selected Compounds/Mappings command from the Tools menu. Or you can double-click on a row number. The conformers are minimized to achieve a fit with the hypothesis, and the result is displayed in the 3D workspace. The mapping can be investigated by rotating the locked objects, using the right mouse button.

6. To look at a different molecule mapped to the hypothesis, repeat the above procedure for a different row in the report. The new selected compound is displayed mapped to the hypothesis.

7. To overlay more than one compound simultaneously on the hypothesis, extend select the row numbers of additional compounds in the report, and repeat the procedure described above. When new compounds are selected, they are superimposed on the others in the workspace in an optimum alignment.

For more information, see "To Display, Delete, Register, and Unregister Individual Conformers."