Bioinformatics and Systems Biology of Aging: Lord of the Numbers
Since January 2016, the group of Hans A. Kestler has been associated to the FLI as a cooperation group between Ulm University and the Leibniz Institute on Aging (FLI) in Jena - located in Ulm.
The rapid development of molecular biology has given rise to an increasing demand in computational and mathematical approaches to analyze and understand the resulting data. In particular, advanced methods from Bioinformatics are required to extract, investigate, and integrate the essential information from high-throughput experiments, such as microarrays or Next Generation Sequencing. The emerging field of Systems Biology provides formal approaches to (temporal) modeling and simulating regulatory processes in biological systems. The research of Hans A. Kestler’s group at the FLI is at the interface of computer science, statistics and life sciences and covers the following main aspects:
- Statistical and data mining approaches for high-throughput data, with an emphasis on feature selection, classification and clustering,
- Modeling, simulating and analysis of regulatory networks, in particular ODE, Boolean, and rule based approaches,
- Visualization and functional annotation.
The advent of high-throughput biomolecular technologies has made high-dimensional biological data available for the investigation of many clinical settings. The large numbers of features and low numbers of probes in such data sets poses many challenges for their analysis. Machine learning approaches and statistical methods are essential for the interpretation of the data. For example, clustering methods can detect groups of similar probes. Feature selection techniques are employed to identify features (e.g. marker genes) that are relevant to distinguish certain phenotypes. Classification algorithms can predict the phenotype of a probe according to the measurements.
Interactions of genes and gene products as well as cross-talks between individual pathways make up complex networks that preclude an intuitive understanding. Along with the increase in knowledge on genetic interactions, mathematical modeling and simulation have become indispensable tools for the analysis of regulatory networks. Modeling approaches vary in the degree of abstraction, which influences the level of detail, but also comprehensibility and the amount of information required to specify the model parameters. Our research covers the highly abstract Boolean models as well as comprehensive models based on differentially equations or intermediate rule based approaches.