This dataset contains important data files for the MD simulation that are part of this publication. The "simulations" directory contains Gromacs parameter files (.mdp) and the run input files (.tpr) as well as the final coordinate files (.gro) of each individual production simulation. The directory "figure3" contains the raw data used to create Figure 3 in the manuscript. The subdirectory "a" contains the data for the PCA projection plot in subfigure 3a. It includes projections of the simulation ensembles of Api88 conformation I-III on to the two dominant conformational modes (.xvg) and the respective extreme conformations (.pdb). The projections of the three initial models and the optimized structure set are also included. Subdirectory "b" contains a numpy array (.npy) with the data for the correlation heatmap in subfigure 3b. Subdirectory "c" contains the results of several correlation-optimization searches. Each directory "N#_maps", where # is to be replaced by the number of structures in the set, contains the search results for N correlation-optimized structures in the Api88 trajectories in the form of a pickled python dictionary (state.pkl). The dictionary has the following keys: used: Already used sets of MD structures (frozenset) selection: Structure set selected in the last iteration (set) weights: weights of each structure in the selected structure set (numpy array) iteration: Counter of the last iteration (int) The directory "supplentary_figure_correlation_time" contains the data for a plot of the optimized correlation coefficient as a function of simulation time. The results of the optimization algorithms (as pickled python objects) are included in the subdirectories with the associated trajectory length as a name.