[ R-Group Decomposition | Free Wilson | SAR Table | Plot R Groups | SALI | Matched Pair Analysis ]
10.33.1 R-Group Decomposition
Tutorial | Video
To decompose a library into fragments based on a Markush scaffold (opposite of R-group (Markush) enumeration ):
- Read the sdf file you wish to decompose into ICM and it will be displayed as a molecular table.
- Sketch the Markush structure you wish to use to decompose the table by and save it in a chemical table.
- Chemistry/SAR Analysis/R-Group Decomposition.
- Choose the table containing the Markush scaffold. Index refers to table row if you have more than one Markush structure in a table.
- Use the drop-down option to select the table you wish to decompose and the column containing the 2D chemical (usually called mol).
- If you have more than one R-group ICM can either generate a different table for each R-group or it can merge it into one single table whereby column will represent R1 and column two R2... This option is useful if you want to generate a SAR table with a column of activity data next to the R1 and R2 columns.
- If you check the box "Auto Add Missing R Groups" then unique R-groups will be extracted from the scaffold where hydrogens can be attached.
- The decomposed table will be reported in a new table with new columns for the R groups (R1,R2...). If you have an additional column with activity data you can sort the table by R-groups and see the effect of an R-Group on activtiy. This table can be used to build an additional SAR table as described here.
10.33.2 Free Wilson Regression Analysis
The free wilson regression analysis can identify interesting combination of substituents which might have been missed by other SAR analysis methods.
- First you need to decompose your hits using a common Markush structure to obtain a set of R-Groups.
- Chemistry/SAR Analysis/Free Wilson Regression Analysis
- Enter the name of your R-Group table
- Choose the activity column.
- The method adds weights into original table R_group_weights and colors them: blue positive contribution to predicted property, red - negativ. Separate tables are returned so it's easier to sort and find what group is most influential in the certain position.
| Tutorial |
To generate a SAR table you first need to decompose the table into the R-Groups.
| Decompose your library. Decompose the library to extract the R-Groups or read in a decomposed SDF file.
| Generate SAR table. Chemistry/SAR Analysis/Generate SAR Tables. One or more columns can be chosen for analysis and multiple tables will be generated. Color the table by a selected column e.g. activity.
| SAR table. The SAR table will be displayed with an activity scale and R-group scaffold in the first cell (A1) and the R groups along each axis. You can use the table to investigate which R-group(s) contribute to activity.
You can make a plot of R-groups from library decomposition. This is described in the plots section of the manual.
10.33.5 Generate SALI Table
| Tutorial |
Structure--activity landscape index: identifying and quantifying activity cliffs.
This method produces a SALI score as described in the publications by John Van Drie and co-workers. It will identify "structure-activity cliffs": pairs of molecules which are most similar but have the largest change in activity.
- Read in a sdf file containing "mol" and "activity" column.
- Chemistry/SAR Analysis/Generate SALI table.
- You can choose to use the Log of the Activity Column.
- The method will try to find pairs based on defined distance cutoff (0.15 is a good threshold for larger datasets).
- A results table will be reported containing the pairs, tanimoto fingerprint similarity and SALI index.
10.33.6 Matched Pair Analysis
| Tutorial |
Matched Pair Analysis (MPA)can be used to study changes in chemical properties based on small well-defined structural modifications to the chemical structure. The method tries to find the Maximum Common Substructure and clusters by similarity. The analysis is performed in two stages: clustering and then Maximum common substructure on each pair in each cluster. The output is similar to SALI output plus it gives the actual R-group pair variation.
The pair table is sorted by score so top hits should show small variations and big score difference. To run MPA:
- Read in a chemical spreadsheet containing a 'mol' column and activity data.
- Chemistry/SAR Analysis/Matched Pair Analysis.
- Enter the name of the Input Table.
- If the table has activity data or some other property use the option use Activity and select the column name from the drop down button.
- Enter a cluster split distance for determining similarity.
- The size of the R-groups for comparison can be set (default is 12 atoms).
- R-groups can be filtered from the analysis. or a custome filter using SMARTs.
- Chirality can be ignored in the comparison if needed.
The results table (tableName_pairs) is fully interactive between the plot and the original table. The hits are sorted by the score column, so top hits should show biggest activity change with smaller groups. The plot is displayed activity 1 versus activity 2 colored and sized by score (red is better).
- ix1 - index number (row number) of first chemical which is hyperlinked to original table.
- mol1 - 2D structure of first chemical.
- R1_1 - group difference between first (ix1) and second chemical (ix2).
- ix2 - index number (row number) of second chemical which is hyperlinked to original table.
- mol2 2D structure of second chemical.
- R1_2 - group difference between second (ix2) and first chemical (ix1).
- scfl - common scaffold
- dvisible ('d' delta 'visible' - delta of your activity column: 'd' + original_column_name)
- score is calculated as normalized dActivity + (1 - normalized weight of both groups).