Résumé : An important goal of biology is to understand the key mechanisms of evolution underlying the diversity of living organisms on Earth. In that respect, the recent innovations in the field of new generation sequencing technologies (NGS) are bringing new and exciting opportunities. This thesis presents results obtained with these tools in the specific context of the study of two sister species of cold-adapted leaf beetles, Gonioctena intermedia and G. quinquepunctata. More specifically, this work is focused around four research directions: the two first explore methods of statistical inference using a spatially explicit model of coalescence, by (1) evaluating the potential of various summary statistics to discriminate phylogeographic hypotheses, and (2) investigating the dispersal abilities of a montane leaf beetle, G. quinquepunctata, using an original method that avoids using summary statistics. The third research direction focuses on the adaptation to cold conditions in this montane leaf beetle, by testing the association between genetic polymorphism across tens of thousands of genetic markers and altitude in samples collected at various elevation levels in the Vosges (France). Finally, the fourth, and last, research axis presents the discovery of mitochondrial heteroplasmy, i.e. the presence in an individual of multiple copies of the mitochondrial genome, in natural populations of G. intermedia. We illustrate, here, how NGS technologies could help identify this phenomenon, probably underestimated in animals, on a large scale.