Embedded Files
In our lab we combine multiple quantitative biology techniques, ranging from fluorescent biosensors, live fluorescence imaging, image analysis, single-cell data analysis and machine learning.
Fluorescent biosensors
We use fluorescent biosensors to monitor ERK and AKT kinase activity in real time, employing kinase translocation reporters (KTRs) for ERK. A nuclear marker is always included to enable accurate segmentation and tracking of individual cells. In the future, we plan to combine these signaling biosensors with reporters for cell cycle progression and apoptosis, allowing us to simultaneously monitor signaling dynamics, cell cycle state, and apoptotic events in the same living cells in real time.
Live microscopy imaging
We perform live microscopy imaging on a variety of cellular models, from 2D epithelia in vitro (MCF10A mammary epithelial cells, melanoma) to 3D patient-derived organoids. We are convinced that observing cellular community in time at single cell resolution allows us to see dynamic and elusive phenomena that cannot be seen with end point measurements. This will open the doors to new observation about cell and tissue biology, both in normal and cancerous tissue.
Image analysis
We combine different image analysis segmentation and tracking modalities to extract biosensor information from individual cells over time. We are constantly interested in implementing new methods for image analysis.
Single-cell temporal data analysis
By combining fluorescent biosensors and image analysis we extract thousands of single cell time series, that reflect changes of kinases activity, or changes in cell state (cell cycle, apoptosis, differentiation). We analyze these time series to explore single-cell signaling dynamics, for instance to identify peaks or other features. Moreover, we identify cluster of coordinated activity such as ERK activity waves.
Machine learning
Exploring the large amount of single-cell data can be cumbersome by human intuition. For this reason we also employ machine learning to explore single-cell data. One example is CODEX, developed by Marc-Antone Jacques in the Pertz lab at the University of Bern. This tool uses a convolutional neural network to explore the single-cell dynamic landscape across multiple conditions (classes).
Page updated
Google Sites
Report abuse