Two-day training and two-hour webinar / November 07, 2018 - November 09, 2018
Training no. 6: Computational toxicology tools
This course will introduce the participants to three different in-silico approaches, namely QSAR models, read-across, and the TTC (threshold of toxicological concern) concept. The course will include demonstrations and explanations of currently available predictive toxicology models such as Derek, CAESAR, TOPKAT, MultiCASE and modules available in the OECD QSAR Application Toolbox. Strengths and weaknesses of individual modeling approaches will be explained. The algorithms and logic behind the available models, model applicability domain and quality issues related to predictions of specific substances (or classes of substances) will be made intuitively understandable. Sampling the underlying databases, looking for toxicologically relevant analogues and subsequently validating a model for a specific type of substances will be illustrated using the OECD QSAR Application Toolbox. A focus will be on the interpretation of results and their reliability, whereas the application of different QSAR models is not a focus of this training.
Read-across, the general principles of similarity, and the different steps of the evaluation will be explained. Recent guidance documents such as the RAAF (ECHA 2017) will be taken into account. Based on read-across examples, the participants will learn about the difficulties in the definition of source compounds and the assessment of uncertainty.
Finally, the TTC concept will be explained: its use and limitations in the absence of experimental data. Special attention will be given to “frequent errors” such as the consideration of exclusion criteria.
This module course consists of four half-day sessions, each a mixture of lectures, discussion and practical (hands-on) sessions. Participants will get an overview of the possibilities of reproducing/validating model predictions provided in dossiers in order to evaluate their usefulness for risk and hazard assessment; and they will get an overview of the most important quality issues making model predictions less reliable or unusable. Hands-on exercises on cases dealing with prioritization of potentially carcinogenic substances (both genotoxic and non-genotoxic) and TTC (threshold of toxicological concern) concepts will give participants a clear idea of the possibilities and difficulties of applying predictive toxicology methods.
During the trainng, the participants will use an EFSA laptop/desktop computer with the OECD QSAR Toolbox software installed in advance, and the hands-on examples will make use of this software. Other tools, such as Lhasa METEOR (metabolism simulation), will be demonstrated in the lectures.
As seating is limited for the two-day training course, some aspects of this topic will be trained in addition in a two-hour webinar held a few days after the two-day training course. You may register for the two-day training course or the webinar, or both. EFSA reserves the right to select the participants of each course from all registered persons. After the registration deadline, all registrants will be informed whether they can actually attend the selected training.
Training materials will be sent to all participants a few days prior to the training/webinar.