Applied Bioinformatics Group

A   A   A
Home > Research > Structural Bioinformatics

Skip to content. | Skip to navigation

Structural Bioinformatics, Biophysics, and Drug Design

— filed under:

In structural bioinformatics we are currently working on methods for protein engineering and design and also develop and apply a broad range of methods for computer-aided drug design. In the latter area, we are actively working on several targets, in targets related to antiinfectives and cancer-related kinases. In biophysics, we are mostly interested in biomolecular energetics, in particular in the understanding of the interaction of proteins, water, and counterions.



Our primary focus lies on the design and implementation of novel docking algorithms for binding mode prediction and also for virtual screening.  Our current docking software BALLDock uses several heuristics like Simulated Annealing and Genetic Algorithms to search for putative binding modes in the search space of all receptor-ligand complexes. The BALLDock is implemented in C++ using our molecular modelling library BALL. We also devised several scoring functions to better estimate the free energy of binding, i.e. for protein-carbohydrate complexes, and are continuing our work in this field to devise other application-specific scoring functions. Docking methods and molecular dynamics simulations are also part of our research in Grid Computing.



QSAR and QSPR methods try to model biochemical properties like toxicity, drug-likeness or binding activity by using Machine Learning methods. The BALL library was equipped with several QSAR modeling methods like PLS, SVR, kNNR that allow the user to build QSAR models using BALL. Further extensions are planned to make QSAR modelling more convenient for BALL users. These methosds are being used in several cooperations with experimental groups for lead finding on numerous targets.


Homology Modeling

Homology Modelling is a concept to predict the tertiary structure of a given target sequence by using known template structures that have high sequence similarity to the target sequence. In our research we use Homology Modelling to get approximate structures of drug targets for docking and binding pocket analysis. We apply these techniques together with docking methods to the identification of possible inhibitors for various targets, e.g., kinases and glycosyl transferases. Part of this research is conducted as part of the DFG-funded SFB 766.

Biophysics Icon  

Theoretical Biophysics

We have long been working on the energetics and dynamics of biomolecular interactions. The interaction of proteins with water and counter ions is one of the most intriguing phenomenons in biophysics. the interactions are complex, nonlocal, and highly dynamic. It is thus little surprising that modeling solvation effects is one of the key hurdles in docking. We are working on new theoretical models to understand these interactions and are closely collaborating with experimental groups interested in studying these effects using scattering techniques.



Over the years we have implemented numerous techniques for molecular modeling and drug design. We have made these methods available through the open-source C++ framework BALL and  BALLView.


People working in this area:

Luis de la Garza, Jens Krüger, Charlotta SchärfePhilipp Thiel, Flávia Passos

Selected publications: