Next to approaches from systems and computational biology, we use data analysis techniques, including machine learning and artificial intelligence, and methods of algorithmic nature for modelling diseases and disease progression. Formalisms like graphs, formal grammars, and automata are used to develop cutting-edge scalable solutions to meet big data challenges. To this end we capitalize on the capabilities of the state-of-the art parallel technology, like GPUs, multi-core processors, and computer clusters.
In close collaboration with hospitals we develop and apply algorithms and models to analyze data of patients and make predictions about treatment. For example, to identify genetic variants in the DNA of patients with inherited metabolic diseases, to develop predictive models to quantify metabolic health in Metabolic Syndrome patients undergoing bariatric surgery, or to extract diagnostic and predictive information from continuous glucose measurements in diabetes patients. Based on molecular data of the patient the models will assist in establishing more accurate diagnosis and more targeted and personalized medication or other therapy. Recent examples in this context include developing diagnostics methods for COVID-19.
We have extensive collaboration with other research groups at BME: BioInterface Science, Biomedical Materials and Chemistry, Molecular Biosensing for Medical Diagnostics, Nanoscopy for Nanomedicine, and Chemical Biology, as well as research groups abroad, like the Institute for Computational Biology (Helmholtz Centre, Munich). We apply advanced data analysis techniques, e.g., deep learning and convolutional neural networks, on a broad range of topics. The examples include influence of topography features of surfaces on cell proliferation, developing of polymer arrays for high throughput screening, topics in digital pathology, like image analysis for automated diagnostic of leukaemia and non-alcoholic fat liver disease.