Extended chemical input and lookup
Compound lookup in web ressources
Files not found for strange reasons
Large model collection crashes
ChemProp data files
ChemProp output: HTML vs. ASCI
Run-time popup warnings
Result export from ChemProp into Excel
ChemProp output appearance
Missing model information in results
Printing of method information
Plain text editing, saving, exporting
Structure vs. substructure search
Edit compounds does not work
Tips for developers
To run ChemProp smoothly, the XP SP3 compatiblity mode should be used. This can be achieved by right-clicking on the icons of the QSAR and database module in the start menu or on the desktop, and selecting properties, then opening the compatibilty tab. The exact wording of these items depends on the language of the operating system. Further adaptations may be required for the database module. If you run into trouble, please contact us. With 64 bit versions (no matter whether XP, Vista, or 7), some particular models do not work. We are working on fixing this.
The ChemProp QSAR module provides only basic chemical input opportunities. To obtain structures from the database, select "Compounds>Compile new molecule file". There, the Database>Basic queries ...(ALT-B) e.g. allows to lookup for names and RN numbers, including a link to eMolecules. Do not forget to move or copy the retrieved molecules into the new compound file provided by this module!
ChemProp provides convenient links to one of the most important chemical WWW ressources (eMolecules). They are available in the database molecule. In a compound list, right-click on a compround provides a menu containing the item "eMolecules". This yields a special SMILES code suitable for a structure search in eMolecules, and directly opens a respective query using this code. In addition, the basic search dialog contains an "eMolecules" button. After typing a valid RN number, this will direct this number to an eMolecules query.
Occasionally, error messages may appear complaining on files not found, even though the files are there
definitely. There may be two reasons:
1. Some older modules written before 2000 rely on short path names. In particular, this affects the internal import from SMILES. The only workaround up to now is to completely install ChemProp in a folder tree with no individual part of the full path name exceeding 8 characters, and to avoid spaces in path/file names. This includes the program files,for that reason, ChemProp is not installed in the usual program folder.
This issue will be fixed in future step by step.
2. User defined files treated in any parameter dialog in the method selections should always be fully quallified, i.e., the full path name is required.
This will be fixed in future by providing a user defined default folder, but files located somewhere else will always require full paths to be found properly.
Furthermore, as with 1., spaces in path and file names should be avoided.
Currently, the number of possible output values per run is limited due to system constraints. There is no fix limit, but depending on the system as well as on the selected models, about 30 to 40 output values are possible, otherwise the run will fail.
So, in case of an error message with a large model collection, please try to split into separate runs first.
The File menu of ChemProp supplies saving and loading of configurations. This can be used to setup several pre-defined sessions. Saving and loading includes options, selected jobs, working directories. ChemProp configuration files (*.ini) can be provided as command line parameters to be loaded at startup. Creating several links to ChemProp with different ini files can be a quite effective way to work with different configurations.
Unfortunately there are still some shortcomings. In particular, in default mode the molecule file is not saved. To achieve saving of the compound file together with the other settings, the user has to save the configuration in a file, and then has to exit ChemProp. Within the Exit dialog, saving of the compound file name can be checked. This procedure is required once for each configuration, after re-starting and re-loading the configuration, the molecule file name will always be stored. For the revers task, the same procedure has to be applied again.
In the method selection dialog, two alternative opportunities to save and load method collections are provided. Save/Load list can be used to save the current job list and to later re-load the list in order to obtain the same job list with individual access to all single jobs, parameter setttings and method information. When re-loading from a file, the job list will complete be replaced by the stored list, i.e. any jobs defined before in the actual session will be removed. In contrast, Save script stores the current joblist as a single job. Re-loading appends the complete task as single job to the current job list, without being able to read or change any setting for that job. Running the calculation will result in the same calculation results and output for that part of the job list. The Save/Load script mechanism is useful to pre-define jobs for fixed tasks.
ChemProp applies a particular file format to deal with data, e.g. to provide experimental data as input or for statistics. This format is quite easy to understand, there is one line for each compound, starting with a line number, and followed by the data columns. Missing data are marked by -999.
ChemProp supports creating of respective data file in the Calculation menu by providing "Run to create data file" and "Run to set for statistics". In both cases, any current option settings violating the format will be ignored. No formatted (HTML) output will appear, but the result may be accessed via the link from the session overview. The resulting data file is the plain text output (to save it for later use, see "Plain text editing ..."). In the second case, the calculation (!) result file additionally will be set to be the experimental (!) data file, and the option dialog for statistics will be opened.
Both opportunities are available only when a compound file is active and the (batch) job list contains one ore more jobs. For statistics, in case of more then one result columns per line the first data column will be applied.
The ASCII output is only included for historical reasons and will be omitted in future version. It is strongly recommended
to uncheck ASCII in the Options dialog, and to check HTML. Checking HTML prevents all additional output from popping up in
separate windows after calculation runs. HTML checked or not, in any case the recommeded access to any output is supplied via the
session overview. It provides links to all output. The session overview can be obtained from Calculation/Results.
This applies to calculation output only. Any editing in expert mode as well as text output from the Compounds menu will still be presented in the external editor. Any HTML output can be directed to the default browser instead of the internal viewer of ChemProp. To do so, check the External HTML browser toggle within the options.
The basic concept of ChemPop is to execute calculations in a batch-like mode, with user intervention at calculation startup only. However, some important warnings as e.g. chemical domain violation may be issued during the run in terms of popup messages. The user may disable this behaviour in the Options dialog by unchecking "Enable popup messages during calculations". In any case, all messages are being logged during the run, and can be accessed later from the overview page.
In Options, set the decimal delimiter to the character Excel expects, set the error / missing data string to any reasonable Excel function string (e.g., in the German Excel version, =nv() is a good choice), and check the HTML toggle, but do not check the External HTML browser toggle.
In the result window, uncheck the "Alternative context menu" toggle!. If you need Excel export regularly, this can be saved as default in the Edit/Preferences dialog. With a proper Excel installation, the right-click context menu then should provide an Export to Excel opportunity. The triggered action depends on the position of the click: With clicking in the result table, the table will be exported solely and immediately. With clicking somewhere outside of the table, an export dialog will appear. E.g., automatic date detection may be toggled off by means of the Options button in the right top corner of this dialog in order to prevent data and CAS numbers from being miss-interpreted.
The majority of the chemprop result output as well as method information is controlled by a HTML style sheet. This sheet is called chemprop.css, and is located in the ChemProp installation subdirectory .../sub/url. It may be modified to fit any demands, needs and tastes. Note: Modifying this file in a network installation will affect the Chemprop appearance for all users!
In a particular setting, the model information may not appear in the report. Both HTML and PDF are effected. It happens when an editor name or path containing spaces has been entered as the external editor in the general options. This bug will nomore be fixed in the current QSAR module, but of course it will not appear in the future unified ChemProp user interface. For now, there is a workaround: Just enclose the entire editor name by quotes.
To avoid displaying and printing of online links, start ChemProp with option -f.
Within the HTML output window, the best way to edit, save, export any plain text results (e.g. details, statistics, etc) is to apply the source editing opportunity of the right-click context menu. This will open the respective file in notepad.
In the Results section of the Options dialog, toggle "Summaries in main and detail output" provides access to column result and descriptor statistics. Once checked, the calculation detail output of fragment tables will be augmented by fragment statistics. Furthermore, the result output (both plain text and formatted) will contain a column statistics summary. In expert mode, this summary will be presented in an alternative (table) version when debug output is checked.
In the database module, the Advanced queries and the External substructure query items provide means to lookup substructures
in molecule files or in the internal or external databases.
The search string will be modified automatically to obtain valid results, even for the typical case of more atoms in the molecule than looked for in the substructure string. For an exactly matching structure search, this needs to be toggled off, in order to obtain "not matching" in case of remaining atoms. It is very easy to extend the search string for that purpose:
- Preceed the string by "once\1 " (The space character is important!)
- Append "\A^|" at the end of the string
This directs the automated modification to a search for the opposite of any atom A^|, which never can match. Since the string is extended to a 2-lines search now, the additional 1 needs to be added according to the specification. Furthermore, once excludes multiple occurrences in larger structures.
From the QSAR module, in some cases the Compounds/Edit presents the empty database modul without showing the compound file. A current work-around is to save the compound file somewhere before using the edit function. Currently, the QSAR functions are being merged with the database module, so this issue will disappear.