 In my group we use artificial intelligence
approaches in various areas. Examples include the use of machine learning
(regressors) to predict affinity of drug candidates (molecules) for drug
targets (proteins) in a polypharmacological context. Input for these models are
properties of the molecule (chemical), properties of the protein (e.g. binding
site), and properties of the interaction (e.g. interaction type). Furthermore we use classifiers in a similar
way to predict activity or toxicity of drug candidates, however for this we
also use images as input (based on high-throughput microscopy). Finally we use
generators to generate novel drug-candidates (using a sequential SMILES format)
that meet one or more predefined criteria. In all these cases we use
experimental validation to validate our models. For this we collaborate heavily
with experimental groups (medicinal chemistry, chemical biology). Recently
we have started exploring the distribution of machine learning models as tools
for chemists. Our models are wrapped and made accessible for chemist to use in
day to day routines as part of their work. We aim to make AI a low threshold
tool that can be used to speed up and improve the classical medicinal
chemistry. Personal
Background My MSc was
in Biopharmaceutical Sciences but during my MSc (upto 2007) and PhD (upto 2013)
I specialized in the application of machine learning in drug discovery. I
obtained my PhD (Leiden University) on a grant sponsored by Janssen
Pharmaceutica. Subsequently I did a postdoc at the European Bioinformatics
Institute in Cambridge (UK) on a Marie Curie / EMBL fellowship. I currently
collaborate with various commercial and non-commercial organizations and within
the EU I am active in both the EUToxRisk and eTRANSAFE IMI consortia. I am always interested in collaborations, you can email me at |
   
  
|
|
