Molecular dynamics (MD) simulation, which employs an empirical potential energy function to describe the interactions between the atoms in a system, is used to investigate atomistic motions of proteins. However, the timescale of many biological processes exceeds the reach of standard MD due to computational limitations. To circumvent these limitations, steered molecular dynamics (SMD), which applies external forces to the simulated system, can be used.Dynamical properties of the gonococcal type IV pilus (GC-T4P) from the bacteria Neisseria gonorrhoeae are first considered. T4 pili are long, filamentous proteins constructed from a subunit (pilin) found to emanate from the surface of pathogenic bacteria. They can withstand large forces (~100 pN), and are implicated in infection. SMD simulations are performed to study the response of the filament to an applied force. Our simulations reveal that stability of the pilus likely results from hydrophobic contacts between pilin domains buried within the filament core. Along the filament surface, gaps are formed between pilin globular head domains. These gaps reveal an amino acid sequence that was also observed to become exposed in the experimentally stretched filament. We propose two other regions initially hidden in the native filament that might become exposed upon stretching.The multidrug resistance transporter EmrD, found in the inner membrane of Escherichia coli is also the target of our studies. EmrD removes harmful drugs from the bacterial cell. We use MD to explore equilibrium dynamics of the protein, and MD/SMD to study drug interactions and transport along its central cavity. Motions supporting a previously proposed lateral diffusion pathway for substrate from the cytoplasmic membrane leaflet into the central cavity were observed. Additionally, interactions of a few specific residues with CCCP have been identified.Finally, we describe network analysis as an approach for analyzing conformational sampling by MD simulations. We demonstrate for several model systems that networks can be used to visualize both the dominant conformational substates of a trajectory and the connectivity between them. Specifically, we compare the results of various clustering algorithms to the network layouts and show how information from both methods can be combined.