Computational medicinal chemistry group


Chris de Graaf


last update: September 2012

Piet Mondrian - Boogie-Woogie

Home

Research

People

Publications

Teaching

Links

COMPUTATIONAL MEDICINAL CHEMISTRY


The Computational Medicinal Chemistry team develops and applies experimentally driven in silico modelling methods for the discovery of bioactive molecules. The integration of the computer-aided discovery and design, synthetic medicinal chemistry, molecular pharmacology, and biochemistry research lines of the Division of Medicinal Chemistry allows the investigation of the structure and function of:
    G protein-coupled receptors
    Ligand-gated ion channels
    Phosphodiesterases
    Kinases
    Cytochrome P450 enzymes


Computatinal Medicinal Chemistry


IN SILICO TOOL DEVELOPMENT


We develop chemoinformatics and in silico chemogenomics tools and protocols that allow explicit incorporation of experimental data:
    ligand-based and protein-based molecular fingerprints (EDprints, FLAP (with MD/University Perugia), EDhotspots.
    protein-ligand interaction fingerprint scoring methods (with CNRS).


In silico Methods


PROTEIN MODELING AND VIRTUAL SCREENING


We have built a track record in protein modelling and virtual screening (VS), exemplified by:
    Winning the international GPCR DOCK 2010 competition to predict the ligand bound crystal structure of the CXCR4 chemokine receptor.
    Obtaining the highest structure-based virtual screening hit rate reported for GPCRs.
    Performing the first successful structure-based virtual screening study to identify small modulators of class B GPCRs (with CNRS).
    Performing the first successful structure-based virtual fragment screening on the GABAA receptor (with (Med)Uni Vienna).
    One of the first studies to predict functional selectivity of GPCR ligands using an agonist customized model based on the antagonists bound GPCR crystal structure (with CNRS).
    Pioneer work in the consideration of water molecules in docking simulations (with ETH-Zurich).
    Our customized protein modeling strategies have been described in several reviews (CYPs, GPCRs).


Modeling VS


In 2009 Dr. Chris de Graaf obtained a personal VENI grant (NWO) to develop selective structure-based virtual screening strategies against intended and undesired protein targets. This has so far resulted in computational models to predict and rationalize ligand selectivity between:
    different receptor/enzyme subtypes (GPCRs, human/parasite PDEs, kinases)
    orthologs (histamine and chemokine GPCRs)
    unrelated proteins (GPCRs/LGICs/CYPs)

We are currently building a fragment-based chemogenomics platform, based on in-house experimental fragment screening data, to construct in silico modeling techniques to increase our understanding of the molecular details of ligand selectivity and cross-pharmacology.


Fragment-based Chemogenomics

2012© Daphné Truan