Some useful hints for developers (Expert mode)
Some useful hints for developers (Expert mode)Substructure search debugging
Restricted substructure search
Atom numbers of search results
In the options menu,
- Set Main output to HTML (only)
- In Detail output, check both (Text ... and Graphical ...)
- In Authoring and trouble shooting, check Debug and Substructures
Now run the job. After finishing, the result window will offer the detailed output and the fragment viewer in the upper right part of the tool bar (or alternatively, as links in the overview page, to be accessed also from there or from Calculation->Results). Further more, the Technical logging will contain the detailed substructure search report. To get the atom numbers this report refers to, use Compounds->View compounds, click at Image options, and check Atom numbers. The view or molecule of interest, or click into an already opened molecule view window to see the atom numbers.
To examine the originally created job list, run the current batch list in debug mode. Then look for the most recent *.psj file in the temporary result folder. This folder can be accessed easily from the result overview.
This file may be loaded again via "Load script" into the batch list.
In some cases it may be required to exclude a particular substructure, however to have this fragments available for later searches. The following trick does the job:
0 structure_to_be_excluded !! :
0 structure_to_be_searched !! :
any_value structure_to_be_searched !! "Any valid marker"Pitfall: Be sure not to have short active during the entire procedure. If necessary, the keyword long may be preceeded, and short may be applied afterward again.
There are occasions, when not only the results of a substructure search are required, but the numbers of the matching atoms in the respective compound file are required.
If there are only a few compounds to investigate, the method to graphically inspect atom numbers as described above (see here: Substructure search debugging) may be sufficient. For a larger compound list, the following procedure to obtain a text list of the respective atoms numbers will be useful:
- Be sure to have the Text window box in the Detail output part of the Options dialog checked.
- In the Methods menu, select Tools / Atom based analysis.
- Open the Parameters dialog.
- Do not forget to select New control file!
- Now, click on the Edit button. Your favorite editor will open with a commented template of the control file. You may either apply this temporary file, or store it somewhere else for later use. In the latter case, the control file selection has to be changed to Existing control file, and the (full path recommended!) file name has to be provided in the respective field (Control file name).
- The control file needs only a few editing steps: At least, in the SUBSTRUCTURE RELATED FILTERING part, the option -s has to be activated by removing the comment sign "#", and the … has to be replaced by the file name of the substructure search table (again, full path).
- Edit the respective substructure search table. For convenience, in the Other files part of the Parameters dialog the file name can be selected (by the File button) or typed in one of the empty fields. This does not affect the job in any way, but provides an Edit button to easily access the substructure table in order to edit it.
- Now, select a compound file and run the method. In the detail output, the matching atoms of each compound will be marked with +1.00000, and the remaining atoms will be denoted as not available. If desired, these marks may be changed in the control file: Fit the -i option in DATA (ATOM BASED VALUES) INPUT for the value of the matching atoms, and/or activate and fit the -u option in the CALCULATION section for the remaining atoms.
- The detail output may further be simplified by activating some or all of the -X options at the end of the control file.
- Typically, substructure search works with a sceleton molecule representation without explicit hydrogen atoms. The HYDROGEN PROCESSING section in the control file may be edited to fine tune the consideration of explicit hydrogen atoms.
- All marked atoms from the entire search table will be found. This may not be sufficient in some cases. It is possible to
split the search into separate runs. For that reason, both the control file and the search table need to be adapted respectively:
- In the control file, uncheck the -t option in the CALCULATION section, and replace … by &Sum.
Define all necessary breaks in the search table by two lines:
- 0 A.^