Team SKI

Protein kinases have been a major target of drug discovery programs for many years due to their central roles in signaling pathways involved in the formation and progression of human cancer, inflammation and Alzheimer’s disease. Most kinase inhibitors that are currently approved and in development bind to the highly conserved ATP-binding pocket of kinases. Given that more than 500 human protein kinases exist, this often leads to low selectivity for the defined target kinase and thus, to undesired side effects or failure in the research and development (R&D) of novel drugs. Optimizing the selectivity profile is therefore a crucial step when it comes to designing novel kinase inhibitors. Moreover, although kinases have been the focus of R&D for more than two decades, a large number of kinases are still unexplored and offer opportunities for novel, competition-free drug discovery projects.

A powerful in silico platform was developed by the Team SKI that guides the design of selective kinase inhibitors. One emphasis and uniqueness of the underlying technologies is that they allow predicting binding profiles across the entire kinome, including so far unexplored kinases and cancer related resistance mutations, and that they can take undesired off-targets into account when designing novel and improved compounds. The developed tools are grounded in computational chemistry and machine learning techniques and include:

• A kinome wide profiling platform for the identification and target interaction analysis of bioactive compounds, trained using a unique panel of biological bioactivity data.
• Multi-dimensional grid approach for the identification of selectivity determining features, capable to take several key and off-targets into account
• De novo design pipeline for lead optimization that generates synthetically accessible compounds with improved activity and selectivity
• Versatile kinome tree viewer (http://kinhub.org/kinmap/).

Publications

1. Volkamer A, Eid S, Turk S, Jaeger S, Rippmann F, Fulle S (2014):
   The pocketome of human kinases: Prioritizing the ATP binding sites of (yet) untapped protein kinases for drug discovery
   Journal of Chemical Information and Modeling 55(3)
2. Volkamer A, Eid S, Turk S, Rippmann F, Fulle S (2016):
   Identification and visualization of kinase-specific subpockets
   Journal of Chemical Information and Modeling 56(2)
3. Merget B, Turk S, Eid S, Rippmann F, Fulle S (2017):
   Profiling prediction of kinase inhibitors: Toward the virtual assay
   Journal of Medicinal Chemistry 60(1)
4. Eid S, Turk S, Volkamer A, Rippmann F, Fulle S (2017):
   KinMap: A web-based tool for interactive navigation through human kinome data
   BMC Bioinformatics 18
5. Turk S, Merget B, Rippmann F, Fulle S (2017):
   Coupling matched molecular pairs with machine learning for virtual compound optimization
   Journal of Chemical Information and Modeling 57(12)
6. Sorgenfrei FA, Fulle S, Merget B (2018):
   Kinome-wide profiling prediction of small molecules
   ChemMedChem 13(6)
7. Jaeger S, Fulle S, Turk S (2018):
   Mol2vec: Unsupervised machine learning approach with chemical intuition
   Journal of Chemical Information and Modeling 58(1)
8. Turk S, Merget B, Eid S, Fulle S (2018):
   From cancer to pain target by automated selectivity inversion of a clinical candidate
   Journal of Medicinal Chemistry 61(11)
9. Schwarz D, Merget B, Deane C, Fulle S (2019):
   Modelling conformational flexibility of kinases in inactive states
   Proteins 87(11)