PD Dr. Oliver Koch
Independent Group Leader
Computational Medicinal Chemistry and Molecular Design
Institute of Pharmaceutical and Medicinal Chemistry
and German Center of Infection Research
Corrensstr. 48, 48149 Münster
Tel.: +49 (0)251 - 83-33443
Fax: +49 (0)251 - 83-32144
Keywords: medicinal chemistry, computational molecular design, cheminformatics, structure-based design, fragment-based design, artifical intelligence, data-driven decision making, design-synthesize-test cycle
The importance of computational methods in pharmaceutical drug research was recently highlighted in a special issue on computer-aided drug design (CADD) strategies in pharma (A CADD-alog of strategies in pharma). The identification and development of new drugs is nowadays hardly imaginable without the use of computational methods. These methods support the complete development workflow from initial target and hit identification up to the development of the final drug candidates.
My research work is generally aimed at developing new bioactive molecules and improving the performance of computational methods in delivering novel and safe small molecule therapeutics. One of the ways in which this is achieved is by gaining a better understanding of protein-ligand interactions. This includes the development of new computational methods, and the extension and application of already existing approaches. The computational work is combined with biochemical evaluation, X-ray crystallography and preparative organic synthesis of small molecular compounds for having the whole design-synthesize-test workflow in one group. The research projects can be summarized as structure-based design and data-driven decision making combined with artificial intelligence towards the development of new bioactive molecules.
On the following pages you can find more information about the group members and our research and scientific output.
03/2021 New Preprint on ChemRxiv available: Natural Product Scores and Fingerprints
Menke, J., Massa, J., Koch, O.* Natural Product Scores and Fingerprints Extracted from Artificial Neural Networks. ChemRxiv. 2021.
03/2021 Successfull Third-Pary Funding: Development of a lecture on "Artificial Intelligence for Pharmacists"
The "Apothekerstiftung Westfalen Lippe" (https://www.akwl.de/apothekerstiftung/) supports the development of a new lecture that deals with artificial intelligence (AI) approaches in cheminformatics and computational drug desgin. Current developments in the field of artificial intelligence (AI) have led to these methods also becoming increasingly important in pharmaceutical research. Therefore, it will be necessary to include AI methods in pharmacy teaching.
03/2021 A new publication online: Computational Ion Channel Research - From the Application of Artificial Intelligence to Molecular Dynamics Simulations
This review provides an overview on a variety of computational methods and software specific to the field of ion-channels. Artificial intelligence (or more precisely machine learning) approaches are applied for the sequence-based prediction of ion channel family, or topology of the transmembrane region. Molecular dynamics simulations combined with computational molecular design methods such as docking and homology modelling can be used for analysing the function of ion channels including ion conductance, different conformational states, binding sites and ligand interactions, and the influence of mutations on their function. Besides highlighting a wide range of successful applications, we also provide a basic introduction to the most important computational methods and discuss best practices to get a rough idea of possible applications and risks.
02/2021 A new publication online: Neural network based fingerprints for virtual screening
Train a fingerprint the important structural features of a target class for enhancing the virtual screening performance. Great work by Janosch now available online. We also provide a kinase specific neural fingerprint for similarity search @ https://github.com/kochgroup/kinase_nnfp. Interesting fact, a graph convolutional network performs worse than an ECFP4 based.