Dresden 2017 – scientific programme
Parts | Days | Selection | Search | Updates | Downloads | Help
BP: Fachverband Biologische Physik
BP 33: Posters - Systems Biology & Gene Expression and Signalling
BP 33.7: Poster
Tuesday, March 21, 2017, 14:00–16:00, P2-OG1
Machine Learning for epigenetic network inference in T cells — •Christoph Kommer1,2,3, Qin Zhang1,2, Ahmed Hegazy4, Max Löhning4,5, and Thomas Höfer1,2 — 1German Cancer Research Center (DKFZ), Heidelberg — 2BioQuant Center, University of Heidelberg — 3HGS MathComp, Institute for Scientific Computing, University of Heidelberg — 4German Rheumatism Research Center, Berlin — 5Charité-Universitätsmedizin, Berlin
T-helper cells direct the cell- and antibody-based arms of the adaptive immune system via the secretion of signalling proteins. The classical view of naïve T-helper cells differentiating into a small number of stable steady states has recently been challenged by experimental findings that point towards multistable hybrid states away from a bistable fixed-point solution.
To interrogate the underlying regulatory mechanisms, we identified the epigenetic landscape in naïve and differentiated T-helper cells from histone modification patterns by applying different machine learning approaches and found distinct classes of enhancers and repressors according to their regulation by lineage-specifying transcription factors and/or extrinsic differentiation signals which are also cell-type specific. For mapping epigenetic states to target genes we furthermore developed a novel parametrized multivariable correlation measure model. With this approach, we recovered well-known cis-regulatory elements and predicted new ones with comparable confidence.
We discuss the utility of these data to learn epigenetic regulatory network topologies in order to explain multistability.