Advancement of cutting edge quantitative technologies

Mathematical modeling – Novel computational approaches are essential to cope with the complexity of the patient-specific dynamic process of cancer formation. We develop mathematical models for the description and the analysis of time-resolved quantitative data obtained on various levels. For the integrative analysis of our multi-OMICS data powerful bioinformatics tools and statistical models are employed. To elucidate mechanisms that control the conversion of extracellular signals into cellular responses such as cell proliferation and differentiation, we utilize coupled ordinary differential equation models of signal transduction pathways and perform parameter estimation based on quantitative experimental data. By means of regression modeling, the link to emergent cellular properties is achieved. These mechanism-based models provide essential building blocks that render agent-based models of tissue organization or pharmacokinetic-pharmacodynamic models of drug-body interactions highly predictive. Our approach enables us to plan most informative experiments, with the aim to reduce the number of animal experiments and to extract best results from limited amounts of patient material.

Parameter estimation of a mathematical model of the IL13-induced JAK2/STAT5 pathway. Symbols: experimental data, solid lines: model simulation, shading: estimated error of the data. Adapted from Raia et al, Cancer Research, 2011.
© dkfz.de

Mass spectrometry – For our data-based mathematical modeling approach high quality quantitative data is essential (Hahn et al., J Proteome Res, 2013; Boehm et al., J Proteome Res, 2014). Mass spectrometry enables us to simultaneously study several thousands of proteins in any biological system. With our Orbitrap mass spectrometers coupled to nano liquid chromatography, we combine label-free and label-based methods such as SILAC or TMT with fractionation methods to determine dynamic changes in the cellular proteome. Additionally, proteome quantifications from scarce patient material allow to determine the abundance of proteins for personalized medicine approaches. Because components involved in signal transduction are typically of low abundance, we are developing a parallel reaction monitoring (PRM) pipeline for their detection.

RNAseq – To study the dynamics of target gene expression we have access to high-end sequencing (Illumina HiSeq 4000) in the DKFZ core facility to perform whole-transcriptome RNA-sequencing. Bioinformatic analysis of the data enables us to identify co-regulated gene clusters and to infer gene regulatory networks.

Live cell imaging – To monitor the activation of signal transduction and cell cycle progression at the single cell level, we utilize live-cell imaging in combination with our own analysis pipeline (Mueller et al., Mol Syst Biol, 2015) that enables us to reliably classify cells in different cell cycle phases and to determine cell death at the cell population level. For the analysis of cell cycle progression, we employ a second-generation reporter mouse model that expresses the fluorescence ubiquitination-based cell cycle indicator (Fucci) consisting on two fluorescently labeled proteins, mCherry-hCdt1 and mVenus-hGeminin. Due to the cell cycle-dependent proteolysis of the fluorescent reporter proteins, the protein levels of Cdt1 and Geminin inversely oscillate thereby making them a suitable indicator of the G1 or S/G2/M cell cycle phases.

Phenotypic readouts – To obtain high-quality data required for mathematical modeling, we advance classical assays for detailed quantitative analyses.For example, we employ automated live-cell imaging and bioinformatic tools to examine cancer cell survival, in-gel invasion, and 2D migration on various substrates. We are generating single-cell tracking data from thousands of cells using self-developed R packages and Image-J macros.

to top