Combi-Chem

4g. Diversity under restraints (C2·LibProfile)

• Property Ranges: to penalize molecules with property values outside specified ranges.

• Property Profiles: to penalize a set of molecules with property distributions (profiles) that deviate from specified profiles.

• Reagent Penalties: to penalize a set of molecules according to number of different reagents, suppliers, cost, etc.

• Select Molecules

  Diverse (Distance-based and Cell-based)

  Similar

  Restrained

• Rgroup Subsetting

  Diverse Library

  Focused Library

  Restrained Library

• Diversity - Penalty.

• Only Diversity (the default).

• Only Penalty (minimize penalty function, disregarding the diversity of the selected molecules).

Note

Specifying property ranges

Accessing the tools

Select COMBI-CHEM I from the list of menu decks, and click LIBRARY ANALYSIS to bring the card forward. Click Restraints Property Ranges to launch the Penalty Ranges control panel.

Using property ranges

1.   Listing the properties for selection. The Get Properties button displays all the numeric properties in the corresponding listbox. The properties may be retrieved either from the Study Table or from a BDF file.

2.   Displaying properties. When you select any property from the listbox, its name, minimum value, maximum value, average, and standard deviation are shown in the panel.

3.   Modifying properties. You can set the desired values for the Lower Bound, the Upper Bound, the Weight, and the standard deviation (Std Dev) of properties you wish to add to the penalty function.

The current property range will be added to the list of active penalties only after the ADD button is pushed. The other buttons allow you to REMOVE the currently selected penalty range from the list of active penalties, to CLEAR all penalties, and to LIST all the penalties to the textport.

4.   Loading and saving. You can save the currently defined penalties to a file by clicking the Save Ranges... button, specifying a file name and clicking SAVE. Similarly, you can load previously saved ranges by clicking the Load Ranges... button, specifying a file name and clicking LOAD.

The file extension for penalty ranges is .prn. The file has a very simple format. There are two lines for each record. The first contains the name of the property and the second lists the Lower Bound, Upper Bound, Weight and Std Dev in order. E.g., for a file containing the three properties, the format would be:

Kappa-2
4.217          13.31           8          1.49236
Kappa-1-AM
7.9          20.44          14          2.2034
Kappa-1
9             23           1          2.56977

Property profiles are loaded from files with the extension .ppf. These files can be created using the Histogram functionality accessible from Preferences/Histograms... on the Study Table (see the QSAR document for more information about using histograms).

Accessing the tools

Select COMBI-CHEM I from the list of menu decks, and click LIBRARY ANALYSIS to bring the card forward. Click Restraints Property Profiles to launch the Penalty Profiles control panel.

Using Property Profiles

1.   Selecting a ppf file. When you select a property profile file in the file browser on the left of the control panel, the properties contained in the file are displayed in the listbox.

2.   Selecting properties. Select the desired property(ies) and push ADD to add the property profile to the list of active penalties. Other buttons are provided to REMOVE specified profiles, CLEAR all profiles, and to LIST all active profiles in the textport.

3.   Printing. To print a profile, select it in the profiles list and click the Print button. The details of the profile will be printed to the textport.

4.   Plotting. To plot a profile, select it in the profiles list and click the Plot button. A histogram plot of the select profile will appear showing the frequency of the property in the library against the value of the property.

MW
6
100.0    150.0    0.10    0.10      1.0
150.0    200.0    0.10    0.10      1.0
200.0    250.0    0.30    0.30      1.0
250.0    300.0    0.30    0.30      1.0
300.0    350.0    0.10    0.10      1.0
350.0    400.0    0.10    0.10      1.0
Hbond acceptor
14
1.0      2.0    0.00    0.00      1.0
2.0      3.0    0.00    0.00      1.0
3.0      4.0    0.10    0.10      1.0
4.0      5.0    0.10    0.10      1.0
5.0      6.0    0.10    0.10      1.0
6.0      7.0    0.10    0.10      1.0
7.0      8.0    0.10    0.10      1.0
8.0      9.0    0.10    0.10      1.0
9.0      10.0   0.10    0.10      1.0
10.0     11.0   0.00    0.00      1.0
11.0     12.0   0.00    0.00      1.0
12.0     13.0   0.10    0.10      1.0
13.0     14.0   0.10    0.10      1.0
14.0     15.0   0.10    0.10      1.0

Specifying reagent penalty restraints

Select COMBI-CHEM I from the list of menu decks, and click LIBRARY ANALYSIS to bring the card forward. Click Restraints Reagent Penalties to launch the Reagent Penalties (LibProfile) control panel.

Reagent Penalties, a type of penalty based on the properties of reagents, is available as part of the penalty functions which bias the selection of diverse (or similar) molecules, including penalties from Property Ranges and from Property Profiles. Any of these penalty terms can be used alone or in combination with other penalties.

The reagents table

• Reagent Name: Unique name for each specific reagent. The name will be used as the key for the lookup to link Rgroup fragment names in the Study Table or BDF file with reagents in the reagents table

• MW: Molecular weight for each reagent.

• Supplier: Name of the specific supplier.

• Cost: Monetary cost of a specific reagent. Cost should be normalized so that all entries in the reagents table refer to cost per unit weight, usually per gram.

• Reagent Penalty: A number with integer values from 0 to 10 representing a preference for a specific reagent. Smaller numbers are better: a reagent with a penalty = 0 is preferred over a reagent with a penalty = 10. This field can be used to penalize reagents based on the presence of particular functional groups or substructures, ease of synthesis, chemist bias or intuition, etc. (data type: integer).

Note

Example reagent table

An example of a reagents table is shown below:

The supplier table

• Supplier Name: a name of a specific supplier. This name will be used as a key to lookup supplier's properties from the reagents table.

• Supplier Penalty: a number with integer values from 0 to 10 to represent a preference for a specific supplier. This field can be used to penalize suppliers based on geographic location, reliability, etc. Suppliers with smaller supplier penalties are preferred.

Note

Example suppliers table

An example of a suppliers table is shown below:

Both the reagents and the suppliers tables are standard .dat ascii files that can be imported into a Cerius2 table for editing, sorting, plotting, etc.

Reagents penalty function

1.   Minimize number of different reagents.

2.   Minimize number of different suppliers.

3.   Minimize total monetary cost of the reagents.

4.   Minimize sum of supplier penalties.

5.   Minimize sum of reagent penalties.

Each of the penalty functions above has an associated user specified penalty. Suppose that of three molecules in the library subset, two use reagent X and the third uses reagent Y. Say the cost of reagent X is 1, and the cost of reagent Y is 2. There are then two types of penalties associated with reagent cost. The first is simply the sum of the costs for each molecule:

P = (1 + 1 + 2) = 4

The second is the sum of the reagent costs for each reagent category, which in this case is simply the cost of reagent X plus the cost or reagent Y, or:

P = (1 + 2) = 3

For more on the reagents penalty function, see Functional form of the reagents penalty function.

Using reagent penalties

1.   Enter tables. Specify Reagents table and Suppliers table by entering the file names into the text boxes or using the Browse... button to load the tables from a directory.

2.   Minimization. Select which features of the tables you wish to minimize: reagent/supplier numbers, reagent costs, and reagent or supplier penalties.

3.   Adjusting the calculation. You may select from three of different options which control how the reagents penalty function is calculated:

a.   When Consider only presence/absence of reagent is checked, each reagent should be counted only once in the penalty calculation (see Theory), rather than summing over all values of reagents.

b.   When the Consider only presence/absence of supplier is checked, each supplier should be counted only once in the penalty calculation (see Theory), rather than summing over all values of suppliers.

c.   When Minimize number of reagents in each Rgroup is checked, the number of unique reagents in each Rgroup is counted and minimized independently, rather than summing and minimizing all reagents in all Rgroups. This is designed to stimulate the reuse of reagents in multi-step synthesis.

Go to the Calculate Penalty control panel (Restraints Calculate Penalties) to specify information printed in the textport when penalties are calculated for selected rows.

List Reagents Penalty Terms: when checked, total Reagent Penalty and penalties corresponding to each of the five terms in the Reagent Penalty Function are listed for every reagent in the currently selected study table rows (or in the selected rows in a BDF file).

List Reagents Quantities/Costs: when checked, the number of times a particular reagent appears in a product, the quantity needed, and the cost is listed for every reagent in the currently selected study table rows (or in the selected rows in a BDF file). The Scale field specifies the quantity to be synthesized of each product, in mmol.

Table 3.

reagent	number of times used	quantity needed (grams)	cost
reagent1	20	10.0	120.9
reagent2	10	12.4	189.2
reagent3	20	22.1	239.3
reagent4	5	7.3	89.3
reagent5	10	12.0	321.6
Total		63.8	960.3

Clear All Penalties