Microbes have successfully colonized the deep subsurface, thanks to their small size and their diverse metabolic activities. This part of the biosphere remains a terra incognita for microbiologists, containing innumerable unknown microbial species and processes. We depend on it for many things, i. e. water supply, oil extraction and nuclear waste disposal. In Switzerland, Opalinus Clay will be used for deep geological repositories, because of its very low permeability. This rock has been studied for 20 years, however, the potential role of microbes has long been overlooked, even though they could play a major role in controlling geochemical conditions. It was showed that they were present in pristine rock, but were not active due to the lack of space. The potential for microbial activity in disturbed rock, under repository relevant conditions, remains unexplored. This is the focus of this thesis. The first condition tested was the presence of space alone. This will occur when the rock is excavated to build the repository. The digging creates a network of large fractures in the rock zones near the galleries. This change is enough to promote microbial activity in a system, mainly composed of two species. The first one, a Peptococcaceae, oxidizes organic carbon to dioxide carbon, by reducing sulfate. The second organism is a Pseudomonas that seems to grow by fermenting organic carbon. From this model, it appears that the Peptococcaceae feeds on low molecular organic acids present in the Opalinus Clay, but also on fermentation products from Pseudomonas. What is less clear is the source of organic carbon for the latter. Is it only feeding on dead microbial cells, or is it also able to feed on reactive fossilized carbon contained in Opalinus Clay? The second case study included an additional energy source, hydrogen. This gas will be produced by anoxic steel corrosion and could damage the repository if the pressure increases significantly. However, it can be oxidized in situ by autotrophic microbial species. The dominating microbe in this system is a sulfate-reducing Desulfobulbaceae. It produces organic carbon from carbon dioxide, which can later feed heterotrophic bacteria, which are also sulfate-reducing and that release carbon dioxide. A carbon loop was thus reconstructed for the first time in the deep subsurface. It includes a fermentation step that releases the organic substrate for heterotrophic sulfate-reducing bacteria that completely oxidize simple organic carbon molecules, i.e., acetate, to carbon dioxide. In this system, hydrogen and sulfate consumption rates are 1.5 and 0.2 µmol·cm-3·day-1, respectively. This means that electrons derived from hydrogen reduce electron acceptors other than sulfate, confirming carbon fixation. These projects highlight the role that microbes can play for the safety of nuclear waste disposal. Hydrogen consumption is a beneficial process, but the overall impact of microbial activity can be negative, because it can promote corrosion and weathering of the different barriers surrounding the waste. Further work is needed in order to obtain a more complete picture of the role played by bacteria in a deep geological repository. Nonetheless, the results presented here represent a large improvement in our understanding of deep subsurface microbiology. Microbial processes were never described in this much detail in these environments, as is done here: from the metabolic pathway level, up to the ecosystem level.