Full contact details
Michael Lynch Laboratory (RC-323)
Regent Court (IS)
I started my academic education with a Bachelor degree in Pharmaceutical Sciences from the University of Freiburg (Germany) and a Master degree in Toxicology from the University of Düsseldorf (Germany). For my Master thesis I investigated with Prof Wilhelm Stahl the inhibition of Cytochrome P450 enzymes by carbon monoxide releasing molecules (CORMs), a potential source for drug-drug interactions of clinical relevance. Subsequently, I moved from experimental to computational research. During a stay at the research group of Dr Andreas Bender in Cambridge, I gained experience in the application of Machine Learning methods for toxicity prediction of chemicals.
Title: "Toxicity Prediction using Artificial Intelligence"
QSAR (quantitative structure-activity relationship) models (i.e. models linking the bioactivity to the molecular structure of a chemical) are a common approach for toxicity prediction across different industries (e.g. medicines, pesticides, cosmetics) as an alternative to expensive and ethically problematic animal studies. However, for some types of toxicity the performance of current approaches is not sufficient to be useful in practice. Also, the lack of interpretability of frequently used algorithms (e.g. Deep Neural Networks) limits the confidence in these models. Currently, I am investigating the potential of multi task QSAR models (i.e. predict several toxic endpoints at the same time) to improve the performance compared to classic single task QSAR models. In another study of my project I will explore possibilities to interpret predictions made by various QSAR models.
Artificial Intelligence, Machine Learning, Toxicology, Pharmacology, Chemistry, Drug Discovery
Funding and awards
PhD scholarship by Lhasa Limited
- Prediction and mechanistic analysis of drug-induced liver injury (DILI) based on chemical structure. Biology Direct, 16(1).
- Effects of frequently applied carbon monoxide releasing molecules (CORMs) in typical CO-sensitive model systems – A comparative in vitro study. Archives of Biochemistry and Biophysics, 687, 108383-108383.
- Carbon monoxide releasing molecule 401 (CORM-401) modulates phase I metabolism of xenobiotics. Toxicology in Vitro, 59, 215-220.