The understanding of biological signaling pathways is of primordial importance both from fundamental and practical points of view. Via a complex network of specifically designed signaling pathways, organisms can sense and react to stimuli in their environment and disturbances in the signaling cascades are involved in a wide range of diseases such as cancer, diabetes and immunodeficiences. In spite of this pivotal role, the molecular details of the signaling events are largely unknown and often experimentally difficult to access.Biological signaling events span many orders of length and time scales, from the electronic/atomic level to the mesoscopic/microscopic domain with accompanying time windows that range from femtoseconds to milliseconds and seconds typical for the full activation of a cascade. A theoretical modeling is thus challenging and necessitates a truly multiscale approach where methods of different fields (quantum mechanical electronic structure calculations, atomistic classical molecular dynamics simulations, coarse grain models and systems biology approaches) have to be combined to gain a comprehensive and realistic picture.In the current proposal, we want to push forward the current frontiers of multiscale modeling approaches by improving the accuracy of the electronic part and by further extending the accessible length and time scales. Specifically, we want to work on the development of improved functionals for DFT and TDDFT, extend our TDDFT-based surface hopping approach with the inclusion of nuclear quantum effects. We also propose the development of a new forcematching procedure for the on-the-fly parameterization of semiempirical electronic structure methods based on DFT/MM reference data that will allow sampling of several orders of magnitude longer time scales in the sampling of the QM region while maintaining essentially DFT accuracy. We will also continue our development in enhanced sampling methods, in particular via the development and testing of a Hamiltonian based replica exchange approach. Furthermore, we will combine the QM/MM and classical simulations with coarse grain models for the simulation of multiprotein aggregates. With the ultimate goal of connecting electronic/atomistic information with the systems level, we are will start the development of a framework for the simulation of entire biological networks. These developments will allow us to investigate electron transfer mediated processes in DNA repair. Furthermore, they will enable us to study the later intermediates of the activation cycle of G protein coupled receptors and the interaction of the activated form with the G proteins.