Cryo-EM has greatly expanded our understanding of the structure and function of many important biological systems. The next challenges now concern the responses to the following questions. When advanced classification techniques of single-particle images fail to resolve highly dynamic regions, how can we reveal the structural heterogeneity hidden in low-resolution areas of cryo-EM maps? How can we optimally integrate cryo-EM with other data to maximize the accuracy of the reconstructed ensembles? And ultimately, how can we squeeze out the most information and best use all the available data? This project aims at answering to these questions through the development of an integrative approach to incorporate cryo-EM data into molecular dynamics simulations and its application to the Type II secretion system.