Machine learning models and Molecular Dynamics (MD) simulations can be used to complement experimental data and proved to be powerful tools to aid to solve and validate structures and dynamics of proteins. By combining these two methods with experimental data, they can prove previously unexplained changes and behaviors in proteins. For my project, I would use both MD simulations and machine learning to sample different conformational changes of photoreceptor proteins and try to correlate my findings to previous obtained experimental data (both x-ray crystallographic scattering and cryo-EM data). Furthermore, the obtained results from the MD simulations can be combined with already existing machine learning models to develop new tools to analyse experimental data.