Simulation of biological processes is an important complement to traditional in vitro and in vivo experimentation, but such simulation often requires enormous computational effort. This is because many biological processes are so-called “rare events", meaning that their timescales are long compared to the timescale of molecular simulation (typically on the order of microseconds). Indeed, using traditional molecular dynamics, the simulation of statistically meaningful numbers of biological rare events is generally completely intractable. Thus, much work has been done to bridge the separation of computational and experimental timescales. Two broad (and non-exhaustive) approaches are (1) enhanced sampling, where the simulations themselves are biased, accelerated, or manipulated to encourage rare events to occur, and (2) dynamical analysis, where existing simulations are statistically recombined to gain information about long-time dynamical statistics from relatively short-time simulations. This thesis aims to apply these approaches to study two proteins: insulin, the hormone that regulates cellular glucose uptake, and KaiB, an element of the core oscillator of the cyanobacterial circadian clock. For all investigated systems, we found an ensemble of pathways that were either energetically or dynamically accessible, highlighting the complexity of biological processes and the folly of assuming a priori that a process of interest follows a single mechanism. The first part of this thesis focuses on the enhanced sampling approach, using a technique named Replica Exchange Umbrella Sampling to characterize the energetic and structural features of the insulin dimer dissociation. The dimer dissociation is an essential prerequisite for cellular insulin binding and ultimate biological function. We discovered a large set of dissociation pathways with comparable free energy barriers, ranging from extremes of conformational selection to induced fit. These results reconciled previous experimental and simulation results, seemingly in disagreement, as part of a broader ensemble of coupled (un)folding and (un)binding. We for the first time explicitly characterized monomeric unfolding during the dissociation, which involved the detachment of insulin's B chain C-terminus from the hydrophobic core of the monomer. We also identified key interfacial rotations, and showed using computational infrared spectroscopy that limiting pathways could be experimentally distinguished based on differences in these rotations and associated backbone amide solvations. The second part of this thesis focuses on the dynamical analysis approach, using a recently-developed method called the Dynamical Galerkin Approximation to explore the dynamics of phenol release from the insulin hexamer. By investigating the mechanisms of phenol release and how they might be altered by targeted insulin mutations, we aimed to understand how one might design new diabetes therapeutics that either encourage or discourage phenol dissociation.We identified and quantitatively characterized six phenol binding/unbinding pathways for wildtype, Ile A10 Val, and Glu B13 Gln mutant insulins. A number of these pathways involved large-scale opening of the primary escape channel, suggesting that the hexamer is much more dynamic than previously appreciated. We show that phenol unbinding is a multipathway process, with no single pathway representing more than 50% of the reactive current and all pathways representing at least 10%. We also showed how the contributions of specific pathways can be manipulated by mutating residues both in and out of the phenolic binding pocket. The final part of the thesis combines dynamical analysis with near-atomic molecular dynamics simulations to investigate mechanisms of fold switching for the cyanobacterial circadian protein KaiB. KaiB can reversibly switch between two stable folds, the so-called “ground state" (gs) and “fold switched state" (fs). We used a combination of near-atomic simulation and dynamical analysis to explore the mechanisms of this fold switch. We compared computational predictions of hydrogen-deuterium exchange to experiments to validate simulation parameters, and added proline isomerization into the model. We further discovered that three prolines (P63, P70, and P71) preferentially occupy the cis state in fs KaiB and the trans state in gs KaiB. The two folds have nearly identical free energies. The mechanisms of fold switching are quite complex. Secondary structure elements in the C-terminal fold switching domain can fold and refold in almost any order, with very little unfolding observed in the otherwise-stable N-terminal domain. The primary free energy barrier is correlated with the breaking of beta sheets in the fold switched state. The isomerization of P63, P70, and P71 largely occurs before this barrier. Overall, the secondary structure elements in the C-terminal fold switching domain act as foldons, tending to fold and unfold as units independent of one another. This work is the first statistical treatment of the mechanisms of fold switching of a metamorphic protein and underscores the inherently multipathway nature of fold switching.