The aim: to build a high-quality prediction computer tool
Hazard and risk screening are in focus for one of the work packages in Mistra SafeChem. Ulf Norinder is developing a computational toolbox to estimate whether chemicals are hazardous or not.
Ulf Norinder, who works at the Department of Computer and Systems Sciences at Stockholm University, coordinates the development and utilization of computer tools in Mistra SafeChem. The task includes gathering information on large sets of compounds to build computer models that can estimate whether chemicals are hazardous or not.
How do you perform your work?
I search for information in different databases and scientific literature, to obtain the necessary information on large numbers of chemical structures. Then I use this to build computer models using advanced mathematical algorithms. This field of research is nowadays well-known outside the scientific community and associated with terms such as machine learning, deep learning and AI, artificial intelligence.
Do you focus on any special substances?
The models are built for many different biological, toxicological, ecotoxicological and environmental targets and these include substances acting on the endocrine (hormone) system as well as substances causing carcinogenicity and concerns for the environment.
What do you hope to achieve?
I hope to be able to build high-quality prediction models that will help both researchers in the work packages and case studies and external partners to characterize substances in products to avoid hazardous chemicals and to be able to choose safer alternatives.
What will be the most important contribution of your work to the programme?
A set of high-quality prediction models to test chemicals and alternative chemicals in new and existing products for toxicological and environmental safety, indicating the status of the present chemicals and if better alternatives seem to be available for consideration.