Daniel Schindler (Universität Potsdam)
Abstract: Amoeboid cell motility is essential for a wide range of biological processes including wound healing, embryonic morphogenesis, and cancer metastasis.
With the ability to generate large quantities of complex multichannel data on cell movement and morphology, there is an increased need for automated and reproducible approaches to extract quantitative data from image sequences.
We introduce a theoretical framework for obtaining smooth representations of the spatial-temporal dynamics during cell movement from stacks of segmented cell contour data. It is based on a combination of Gaussian process regression and reproducing kernel Hilbert spaces. Thereon based, we present the Optimal Contour Flow Method (OCFM), a novel approach to continuously track points of reference between successive cell contours, so-called virtual markers, in space and time.
The main advantages over existing approaches is its ability to track virtual markers over the entire observation time span. This allows to study characteristics of cell membrane deformations, like the speed of protrusions or the change in local curvature, and their correlation to protein patterns near the cell membrane in time. In addition, it allows for a new type of graphical representation of the data, extending commonly used kymographs.
We used time-laps microscopy data of the social amoeba Dictyostelium discoideum to illustrate our OCFM approach.
You can also find some information about the Ph.D. activities here.
