Providing the data and software services within OpenRiskNet virtual environment and make the functionality available via application programming interfaces (APIs) already allows for the development of advanced workflows providing solutions to complex risk assessment tasks (demonstrated in the case studies webinars). However, identification and selection of the most appropriate service, understanding their functionality and format the input and output still need manual investigation and programming/scripting skills. To be able to automate these tasks, OpenRiskNet is exploring semantic annotation on different levels.

After a short introduction to ontologies applicable to service annotation, we will first demonstrate high-level annotation, which gives the information for the OpenRiskNet registry service to index the service and to make it findable for users and gives basic categorisation like dataset, algorithm and model. Then we will show low level annotation, which will be used for dataset and input/output harmonisation so that the user but more important other services can understand the information provided/required and how to access it. Finally, first approaches to define the middle layer providing metadata describing the service in more detail like available toxicity domains, available endpoint as well as gene/pathway/AOP associations will be shown.