computational
modeling in biology

Schriftgröße »A . A+ . A++ .

Regulatory networks

In the context of Systems Biology the CMB group analyses and predicts regulatory networks comprising various biological entities. These graph based approaches allow for the integration of heterogeneous data, from single cell to genome-wide expression profiles. We generate regulatory models for the prediction and investigation of network interactions and network dynamics.

Gene Regulatory Network (GRN) inference based on Bayesian Statistics

  • We estimate GRN based on time dependent or steady state gene expression profiles (mRNA and miRNA).
  • In general we here concentrate on semi-scale network sizes (about 10 to 15 genes). Edges in our networks refer to gene-gene interaction, but can also denote indirect interactions.
  • To deal with common optimization issues like overfitting or parameter indeterminacies we here usa an approach based on Bayesian statistics. Furthermore, this allows us to easily integrate prior knowledge.
  • Using a sampling method based on MCMC methods we generate an ensemble of networks that allows us to estimate parameter probabilities and dependencies.
  • Here we strongly collaborate with the groups of Heiko Lickert and Magdalena Götz within the CoReNe project.

GraDe (graph-decorrelation algorithm)

  • Based on the time-delayed correlation for time series data we define a graph-delayed correlation for data with network like dependency structure, e.g. measurements of players in regulatory networks. We use this definition in the context of blind source separation and assume different sources to have vanishing graph-delayed correlation and therefore represent different biological processes.
  • A first realization of this idea was the GraDe Algorithm (graph-decorrelation algorithm) as a sample based approach. The current work exists of a probabilistic formulation with the adventage of the flexibility of bayesian statistics. We can easily include prior knowledge and use the principle of automatic relevance determination to decide which sources are the interesting ones.

Gene Regulatory Networks controlling skin aging

  • We are part of the GerontoSys consortium which aims to infer a gene-oriented stromal aging model that is able to explain regulatory mechanisms on different levels (mRNA, microRNA and proteomics)
  • The application of biostatistics and machine learning approaches for the integration of heterogeneous data enables us to combine transcriptional and post-transcriptional regulation within gene regulatory networks 
  • In addition, we closely collaborate with the group of Prof.Reifenberger at the Heinrich Heine University Duesseldorf in order to reveal the impact of microRNA regulation in UV-induced and intrinsic skin aging

miTALOS

  • MicroRNAs regulate cellular signal transduction via controlling mRNA levels and thus tuning protein abundance. miTALOS analyzes their functional impact by mapping predicted microRNA targets onto signaling pathways. It uses a standard enrichment method as well as a novel proximity score that takes network structure into account.
  • As an additional feature, miTALOS considers the tissue-specific expression signatures of microRNAs and target transcripts to improve the analysis of microRNA regulation in biological pathways. Multiple microRNAs can be combined in a single analysis to highlight their combinatorial effects. A graphical visualization of microRNA targets is provided to illustrate their respective pathway context.

References

D. Lutter, C. Marr, J. Krumsiek, E. Lang and FJ. Theis. Intronic microRNAs support their host genes by mediating synergistic and antagonistic regulatory effects. BMC Genomics, 11(224), 2010. 10.1186/1471-2164-11-224. [ DOI | PubMed | .pdf ]

A. Kowarsch, F. Blöchl, S. Bohl, M. Saile, N. Gretz, U. Klingmüller and F. Theis. Knowledge-based matrix factorization temporally resolves the cellular responses to IL-6 stimulation. BMC Bioinformatics, 11(585), 2010. 10.1186/1471-2105-11-585. [ DOI | PubMed | .pdf ]

A. Kowarsch, M. Preusse, C. Marr and F. Theis. miTALOS: analyzing the tissue-specific regulation of signaling pathways by human and mouse microRNAs. RNA, 17(809-819), 2011. 10.1261/rna.2474511. [ DOI | PubMed ]