QSAR Toolbox: Training Course Agenda

Detailed agenda
Basic course

Basic definitions and general functionalities
At the beginning of the course we cover the core features and terminology of QSAR Toolbox, including document and endpoint trees, data matrix, profilers, target endpoint and more. Explore each of the Toolbox modules - Input, Profiling, Data, Category definition, Data Gap Filling, Reporting and learn the possibilities they offer.
Back to agenda

Example case studies on predicting ecotoxicological and environmental fate hazards endpoints
In this section, the principles of building category approach hypothesis will be demonstrated with several case studies. Our goals are: (1) to introduce the main information which could be obtained in each of the modules in the system and (2) to demonstrate the process of building and refining chemical categories.
Back to agenda

Alert performance and its application
A very simple and useful functionality is available in QSAR Toolbox, allowing to estimate performance of functional groups and alerts with respect to a target endpoint. The information obtained from this functionality prove to be crucial in various practical scenarios, such as: (1) evaluating reliability of identified alerts for read across applications; (2) comparing alerts in multifunctional chemicals based on their performance to resolve cases when there are no analogues; (3) identifying alerts with low performance illustrating their inadequacy for searching analogues and read across prediction purposes.
Back to agenda

Empirical structural similarity
The main purpose of this topic is to fully understand the importance of accurately applying the calculated empirical structural similarity. When and how to apply it effectively. You will learn about the settings that influence the structural similarity estimation, including the method of estimating similarity according to Tanimoto, Dice, etc., the equation and molecular features used for their calculation. The different measures and setting demonstrate the operation character of similarity. This will be demonstrated with example chemicals.
Back to agenda

Prediction report
Report of the prediction follows simple templates which are developed in accordance with ECHA requirements. Specific sections of the report will be demonstrated related to reporting AMES mutagenicity data according to TG 471.
Back to agenda

Grouping chemicals based on custom criteria
Functionalities allowing to define a chemical category based on custom criteria such as user-defined fragment or phys-chem property will be introduced. The goal of this topic is to demonstrate how to collect analogues or refine a chemical group in case of lacking appropriate profiler for the endpoint of interest. This will be demonstrated with case studies.
Back to agenda

Metabolism – simulation and use: Introduction
Consideration of the metabolic activation is an essential component of every toxicological assessment concerning especially human health hazards. In general, the usage of metabolism for read across purposes is a key component of the advanced training. In this context, we introduce straightforward examples to illustrate the importance of metabolic activation in the toxicological evaluation of chemicals and for building the read across hypothesis. Two case studies will be presented demonstrating the assessment of skin sensitization accounting for the metabolic activation.
Back to agenda

Automated and standardized workflows
Manual read across using the information in the Toolbox sometimes is a time-consuming process, isn’t it? What if instead of manual work, the data gaps could be automatically filled in by read-across? Would you say it sounds amazing? The good news is that you'll find out about that very option in Toolbox right here. Our main goal is to present the conceptual algorithms used to build the so-called automated workflows for predicting skin sensitization and acute aquatic toxicity and exemplify them. Key component here is the development of an automated workflow for the purposes of OECD TG 497 which is associated with the defined approaches on skin sensitization (DASS). The DASS automated workflow is in QSAR Toolbox and will be introduced and demonstrated with a few case studies.
Back to agenda

Export of data
The system provides a lot of functionalities to export results from QSAR Toolbox, including experimental and predicted data, profiling results, calculated 2D/3D parameters. Additional settings in the exporting functionality allows to filter and select specific information which to be exported.
Back to agenda

Toolbox Repository
Are you aware that QSAR Toolbox has its own repository and its own collection of plug-ins? It could be accessed directly from QSAR Toolbox or through internet browser, allowing to select and install the plug-ins you are interested in. One of them is the unlocking plug-in of ECHA REACH database which will allow to export data from ECHA REACH database (otherwise you are not allowed to do it).
Back to agenda

Advanced course 1

Cramer profiling scheme
The Cramer classification scheme is widely used for assessing the toxicological profiles of chemicals. In QSAR Toolbox, the latest version goes beyond the original approach, incorporating key expansions and updates developed in collaboration with experts from the EFSA Working Group on Flavourings. Explore the enhanced, comprehensive reporting that provides detailed insights into the structural features of your target chemical, along with a clear, transparent profiling pathway—supporting more informed and reliable decision-making and strengthen regulatory acceptance with confidence.
Back to agenda

Building knowledge platform and usage
You know what a Toolbox profiler is—but do you know how to create your own? Learn how to build custom profilers using your expertise and apply them effectively for (sub)categorization and screening purposes. Explore practical examples, including building a profiler to identify precursors of challenging metabolites like non-degradable or known toxic substances.
Back to agenda

(Q)SAR models in QSAR Toolbox
QSAR Toolbox offers a rich library of ready-to-use well known (Q)SAR models for various properties (such as the ECOSAR and Danish EPA models) —learn how to apply them to predict your target chemicals. Go a step further by discovering how to build and customize your own (Q)SAR models, giving you greater flexibility and control in your assessments.
Back to agenda

Toolbox Repository
Want to expand your Toolbox capabilities with publicly available modules? Discover the Toolbox repository, explore a wide range of profilers, (Q)SAR models and databases and learn how to easily upload and integrate them into your installation. Unlock new functionalities and make the most of the tools available to you. Strengthen your read-across hypothesis by effectively integrating results from VEGA, KATE, OPERA and other models directly within the Toolbox.
Back to agenda

Workflow editor
A powerful new feature in Toolbox that lets you design and implement your own decision schemes for screening and/or prediction. Learn how to build automated workflows that follow your logic step by step—saving time, increasing consistency, and streamlining your entire analysis process. Practice with a simple example to automatically find chemicals sharing the same mechanisms and organic functional groups.
Back to agenda

Custom calculators
Learn how to create your own calculator for predicting 2D and 3D properties using experimental data (e.g. for logKow). You’ll also discover how to apply it as a descriptor in the data gap filling process—enhancing the accuracy and reliability of your predictions.
Back to agenda

Import/export of data
Learn how to import your custom databases into QSAR Toolbox, including seamless integration with IUCLID. Take advantage of Toolbox–IUCLID interoperability to transfer your Toolbox predictions directly into your IUCLID database, and strengthen them in line with QAF criteria. You’ll also discover how to enrich your assessments by importing additional toxicological and metabolism data from the Toolbox Repository.
Back to agenda

Handling of mixtures
Learn how to define your mixtures in QSAR Toolbox and how to specify the exact quantities of each component. Discover how to predict mixture properties based on whether components share a similar mode of action or act independently and generate the relevant specific Toolbox report —helping you to make more accurate and scientifically sound assessments. Examples with acute aquatic toxicity and skin sensitization will be demonstrated.
Back to agenda

Toolbox reports: justification documents
Explore the dedicated justification documents generated alongside the main Toolbox report—designed to support your regulatory submissions with confidence. These documents incorporate the relevant RAAF/QAF criteria, tailored to the specific type of prediction, ensuring your assessments are transparent, consistent, and aligned with regulatory expectations. See practical examples on reporting cases for analogue approaches and (Q)SAR modeling results.
Back to agenda

Search in IUCLID databases. Composition search
Unable to search directly in your IUCLID database? Once your data is transferred to QSAR Toolbox, you can leverage its powerful search capabilities. Easily search for substances, data, and even explore substance compositions for specific impurities and/or additives, e.g. searching of substances with known impurity having isocyanate fragment. Use the simple search engine, organized like IUCLID endpoints, or take advantage of the advanced IUCLID search to define detailed, precise search criteria.
Back to agenda

Query Tool functionality
The Query tool allow you to perform strategic searches across its databases for chemicals and/or data. Learn all the available search criteria (such as structural, mechenistic, similarity, etc) and how to combine them logically—for example, identify chemicals that are Ames-positive but non-carcinogenic according to experimental data—enabling precise, targeted data mining for your assessments and scientific work.
Back to agenda

Endpoint vs. endpoint correlations
Learn how to correlate different endpoints using chemicals with available experimental data. Apply these correlations to explore whether in vivo endpoints (e.g., EC3 LLNA) can be predicted from in vitro data (e.g., EC3 KeratinoSens), enhancing your ability to fill data gaps efficiently by saving time and reduce the cost associated with animal testing.
Back to agenda

WebClient – searching chemicals and data
Access QSAR Toolbox directly from your web browser and explore its WebClient interface. Learn how to efficiently search for chemicals and data, and discover the full range of features and capabilities the WebClient offers for a flexible and streamlined user experience.
Back to agenda

Advanced course 2

Metabolism use
Learn about the available metabolism simulators and observed metabolism databases within QSAR Toolbox what type of information they provide. Discover the different scenarios in which this information can be applied to identify suitable analogues for read-across purposes, such as searching for analogues sharing the same metabolic pattern or the same active metabolite, refining a chemical category by excluding chemicals with different metabolic activation, predicting substances using data for their active metabolite, and more. Each scenario is illustrated with practical case studies, helping you understand how to use metabolism data effectively in real-world assessments.
Back to agenda

Observed metabolism data
Explore databases with observed metabolism data in QSAR Toolbox and gain detailed insights into documented metabolic maps and reported quantities of metabolites. Learn how to apply specific filters to focus on metabolites of interest (e.g. rat metabolites, found in urine), enabling more targeted analysis.
Back to agenda

Alert performance and its application – Part II
You already know what alert performance is, but in this second part you will take it further. Learn how to check the predictability of the alerts identified as a result of metabolic activation and find out how to select the most appropriate alert for searching analogues when multiple mechanisms after metabolism.
Back to agenda

Category consistency
Discover how the consistency of your chemical category can vary depending on the target endpoint. Evaluate your category by analyzing key aspects of similarity—physicochemical, structural, mechanistic, and ADME properties. Build stronger weight of evidence and develop clear, well-supported justifications to enhance your reporting and support regulatory submissions.
Back to agenda

Toolbox reports: justification documents
Refresh your knowledge of Read-across assessment framework (RAAF) scenarios and the associated assessment elements through hands-on training focused on category approaches for read-across. Compare requirements for ecotoxicological vs. human health endpoints and learn through selected practical case studies.
Back to agenda

Predicting higher tier endpoints (HTEs)
Take on the most challenging cases in chemical assessment and learn how to approach complex endpoints for prediction. Discover the data and knowledge sources available in Toolbox for higher-tier endpoints, including repeated dose toxicity, developmental and reproductive toxicity, and carcinogenicity. Learn how to analyse the observed data for the analogues. Understand the key modelling concepts, and explore strategies for different cases including cases where analogues are limited or unavailable.

Analyse questions like:
What is the quantity of the parent chemical after metabolism and is it enough to cause effect?
Are the (a)biotic transformations of the parent chemical fast?
Which of the transformation products need to be considered in the analysis and what are their quantities?
In case of toxic metabolite(s), how to estimate the dose of the parent chemical which will not lead to generation of these metabolites in an amount to cause harmful effects?
Is it possible to streamline the entire workflow for predicting HTEs?

Find out also how to define other relevant endpoints to support the hypothesis for the main target endpoint and how to collect enough weight of evidences for more reliable predictions.
Back to agenda