"Photolyases are natural proteins that by absorbing sunlight can repair human DNA photo-damage. In spite of great potential for use of photolyases in future sunscreen techniques for skin cancer prevention, the detailed mechanism of their operation still has not been established. In this project we aim to unravel their DNA photo-repair mechanism in atomistic details. In order to deliver upon this promise, it must be possible to accurately predict, not only electronic states of isolated active centers, but also their nuclear dynamics after photo-excitation, and how this dynamics is changed through the interaction of the active centers with the protein environment.The objectives of the current proposal is to go beyond a static mechanistic picture obtain from reaction paths developing along active center potential energy surfaces. Our target is to advance our knowledge on the non-adiabatic dynamics of photochemical processes and the effects of an environment upon them, but also to consolidate methodology and procedures that allow predictability and transferability of simulation results. This will be done by systematic comparison and assessment of different simulation methods at distinct levels of theory. We will start with low dimensional parametrizations of potential energy surfaces obtained from accurate electronic structure calculations. These parametrizations will be used in quantum dynamical simulations and will be further augmented by the ""environmental"" effects through the QM/MM approach and system-bath projection formalisms.This proposal builds upon the highly complementary skills of the researcher in potential energy fitting and non-adiabatic dynamics, and expertise of the hosts in non-equilibrium quantum statistics, QM/MM, and electronic structure methods. The training provided by the fellowship will empower the researcher with a very complete set of tools that will be instrumental in achieving professional maturity."