Deep geological repositories in clay formations are a promising option to ensure the long-term isolation of nuclear waste from the population and the environment. In Switzerland, the Opalinus Clay (OPA) formation is a shale whose favorable barrier properties have been characterized in the framework of the Mont Terri Laboratory project. Yet, faults intersecting the formation brings the integrity of future repositories into question. Faults, indeed, might reactivate upon any physical hydro-mechanical perturbation resulting in induced seismicity or creation of preferential paths for fluid leakage. How these fault arrays will reactivate, i.e., aseismic or seismic, and whether dilatancy or compaction will accompany reactivation are far from being well established, yet are of paramount importance to furthering any predictive capabilities. In this context, the objective of this research is to study, through laboratory experiments, the frictional and transport properties of the fault zones intersecting the OPA formation at relevant conditions for nuclear waste storage. The first section of this study reveals a fault gouge that has a pore network dominated by nanopores, yet a higher porosity, and slightly higher permeability with respect to the surrounding non-deformed rock. Furthermore, analyses show a lack of calcite content within the fault gouge, in agreement with recent evidence suggesting pore fluids flowing throughout it. Based on these results, the fault gouge does not act as a barrier; rather it can act as preferential but localized and narrow fluid flow channel favoring fluid transportation. The second part of this research reveals 1) a weak frictional strength of the OPA fault gouge, however extremely weaker under wet conditions, 2) a clear aseismic stable behavior for wet and partially saturated samples, yet a transition from unstable to stable behavior with increasing sliding velocity for dry samples, 3) almost null frictional healing, i.e., a lack of re-strengthening during interseismic periods, 4) cataclastic deformation processes and, 5) on wet experiments, shear-enhanced dilation and a small increase in permeability after shearing. All these results indicate that OPA fault gouge could be easily reactivated via aseismic creep, possibly acting as weak fluid conduits. However, if temporarily dried, they could become potentially unstable. A final section of this work presents the frictional response of simulated scaly clays. Simulated and natural scaly clay fabrics present significant similarities, notably mirror-like surfaces which are formed at sub-seismic velocities and low normal stresses in the laboratory. The simulated scaly fabrics exhibit 1) a lower frictional strength than the fault gouge at same partially saturated conditions, 2) both stable and unstable behavior, i.e., the co-existence of velocity-strengthening and weakening slip patches, and 3) low frictional healing. These observations suggest that the scaly clay fabrics are prone to host earthquakes, yet they might be small and rare over time. This Ph.D. work finds direct implications for the concept of deep geological repositories in clays. In spite of favorable barrier properties of non-deformed OPA, faults cannot be ignored. Hence, this study might be the starting point for the long-term risk mitigation strategies.