Berlin 2018 – scientific programme

Parts | Days | Selection | Search | Updates | Downloads | Help

BP: Fachverband Biologische Physik

BP 10: Postersession II

BP 10.58: Poster

Monday, March 12, 2018, 17:30–19:30, Poster C

Parameter-free high-resolution traction force microscopy — •Yunfei Huang, Gerhard Gompper, and Benedikt Sabass — Institute of Complex Systems 2, Forschungszentrum Juelich , 52425 Juelich, Germany

In Traction Force Microscopy, elastic displacements caused by mechanical interaction of cells with their environment are employed to calculate cellular traction forces. Calculation of traction from displacement is a linear problem. However, as a result of insufficient measurement density and a long-range interaction kernel, the calculated traction values strongly depend on the chosen method for solving the problem. Here, we systematically test the performance of state-of-the-art methods from sparse learning, computer vision, and Bayesian inference for traction force microscopy. Classical approaches include L2- and L1-regularization or spatial filters that depend on an ad-hoc choice of parameters. We also study three further parameter-dependent approaches, namely Elastic Net (EN) regularization, Proximal Gradient Lasso (PGL), and Proximal Gradient Elastic Net (PGEN). Next, we pioneer the use of parameter-free methods such as Bayesian Compressive Sensing (BCS), Bayesian Lasso (BL), and Bayesian Elastic Net (BEN). We introduced novel computational methods for traction force microscopy that eliminate the need for user-defined filter-parameters and also exhibit excellent performance with regard to resolution and accuracy. These methods can enable an objective detection and quantitative measurement of forces at minute cellular adhesion sites.