1 of 1 people found this helpful
Please find below short explanations of the Toolbox modules:
1. Input - allows to enter a target chemical (by CAS, Name, drawing) or a list of chemicals
2. Profiling - Here you can select more than one profiler at once. The profiling results provides information for the structural features and possible mechanisms. The outcome of the profiling also determines the most appropriate way to search for analogues. The profiling results are not predictions and should not be used as such.
3. Data - All databases (containing chemicals with experimental data) and inventories (containing chemicals without experimental data) are included in this module. You can select one or more than one databases. These databases will be used for searching of analogues.
4. Category definition - here will see the same profilers as those in the Profiling module. You should select one of them (based on the outcome of the Profiling module) to search for analogues.
5. Data Gap Filling - Once the analogues are found and data is gathered, you can apply some of the gap filling methods (read across, trend analysis, etc.) depending on the endpoint of interest
6. Report - once the prediction is done, it could be reported. The report wizard contains sections which are automatically populated by the system as well as sections that need to be filled by the user.
In the Input section you can define your target endpoint (example with in vitro mutagenicity is shown below). Once the endpoint is defined, the system will highlighted only the profilers/databases relevant your endpoint.
Hope this is of help.
Thank you for your answer. Some of the doubts that i had are clear now.
I would like to ask one more question. There are cases where we have to group the metabolites based on their structure similarity. Sometimes there are cases where one active substance has more than 40 metabolites. Is it possible to add more than one structures in the "input" and check the structure similarity ?
In the Input module you can enter a single target chemical or a chemical list (Figure 1). Using the buttons in the Chemical List section you can load all of the chemicals from any of the available Toolbox databases/inventories or to load a custom list (your own file). The custom list could be in .txt, .smi, .sdf or .mol format.
Once the list with chemicals is loaded, go to the Profiling module. Find Structural similarity profile (1) and go to the options in the right-click menu (2) (Figure 2). Click Define button (3). 2D editor will appear where you can draw or paste the SMILES of the structure of interest (the structural similarity will be calculated with respect to this structure). Confirm by OK in the editor window and in the Options window (4).
Then click on the Apply button to see the similarity between the specified structure and all structures on the matrix.
In case you want to check the similarity between the specified target structure and the metabolites of the chemicals on the matrix, you need to check also any of the simulators and then to click Apply (Figure 3). The results will appear on the data matrix in the Profile >> Metabolism/Transformation section. You will see the similarity result for each of the metabolites produced by the selected simulator.*
In order to group the metabolites according to their similarity to the target structure, apply right click over the Structure similarity row and select Profile statistic option (1) (Figure 4). Check Group by category (2) and select a similarity range (3). Then you can save the structures in a .smi file or to continue working with them by adding the structures in a new file (4).
*Please note that profiling/metabolizing of a custom file will take some time when you run the chemicals for the first time. Once the chemicals are profiled, their results will be cached.
Hope this is of help.
Thanks for explaning it in so much detail. You have reduced a big time consuming steps in the OECD toolbox for me
I have some more questions: How we can perform QSAR analysis with this tool box? It is mentioned in one of the tutorial of OECD toolbox (i dont know which one) that QSAR assessment cannot be performed with this toolbox, only read across assessment can be performed. I asked the same question to EFSA directly in a conference and the lady from EFSA told me that it was a mistake to name this toolbox as QSAR toolbox. But here the discussion forum tells a different story as you can see there are alot of posts related to QSAR. Could you explain it briefly how we can use OECD toolbox for QSAR assessments ?
My second question is that is it possible to perform QSAR/read across assessment to calculate the LD50 for the endpoint oral toxicity, and what would be the suitable profiler in the category definition for oral toxicity ?
To your questions:
1) Could you explain it briefly how we can use OECD toolbox for QSAR assessments ?
The OECD QSAR Toolbox is not a single (Q)SAR. The OECD QSAR Toolbox is a decision-support tool providing scientific computational methods to find experimental data for chemicals, group substances into categories based on chemical and mechanistic similarity, and predict substance properties without testing.
By making use of the system, hazard and risk assessors are able to:
- Use predefined categories, or to refine existing or build new categories.
- Identify analogous chemicals (or category) based on user selected characteristics. Categorize chemicals accounting for their metabolism: rate of disappearance, formation of stable metabolites, formation of high reactive intermediates, deactivation pathways, etc.
- Extract all available experimental or pre-calculated data from local and remote (web) based databases accompanied with information about their reliability: experimental error, analytical or computational method used, replicates, etc.
- Fill the gaps of missing information within the category by making use of chemometrics approaches such as read across, trend analysis, and (Q)SAR models.
QSAR predictions are accompanied with information concerning their mechanistic background, training chemicals, statistics, applicability domain and validity.
2) Is it possible to perform QSAR/read across assessment to calculate the LD50 for the endpoint oral toxicity, and what would be the suitable profiler in the category definition for oral toxicity?
Yes, it is possible. Please find below an example on how to do this.
- Enter your target chemical by CAS or SMILES
- Define the target endpoint – once the endpoint is defined, the Toolbox will highlight the endpoints which are suitable to be used and the databases that contains such kind of information. For LD50 you have to define the endpoint like this:
3. Go to the Data module and check the highlighted database only
4. In Category Definition module you can use some of the suitable (highlighted profilers) for searching of analogues, e.g. US-EPA New Chemical categories or Organic functional groups profilers.
5. Once the analogues are found, go to the Data Gap Filling stage and click Read-Across.
6. Apply some subcategorizations in order to reduce the uncertainties between the analogues (in the right-side panel: Select/filter data >> Subcategorize)
7. Accept the prediction
8. Go to the Report module and click Prediction in order to generate a prediction report.
More information for each of the Toolbox functionalities could be seen by pressing F1 button of the keyboard.
Tutorials regarding different endpoints and Toolbox functionalities could be found here: https://qsartoolbox.freshdesk.com/solution/folders/16000085980
Answers on other FAQ could be found here: https://qsartoolbox.freshdesk.com/solution/folders/16000083350
Additionally, there are regular annual training courses (two level training: Basic and Advanced) on QSAR Toolbox that take place twice every year in Barcelona - http://www.reachmonitor.org/index.php?lang=2&aptd=4&id=73