As the value of the longitudinal momentum carried partons in a ultra-relativistic hadron becomes small, one observes a growth of their density. When the parton density becomes close to a value of order $1/alpha_s$, it does not grow any longer, it saturates. These high density effects seem to be well described by the Color Glass Condensate effective field theory. On the experimental side, the LHC provides the best tool ever for reaching the saturated phase of hadronic matter. For this reason saturation physics is a very active branch of QCD during these past and coming years since saturation theories and experimental data can be compared. I first deal with the phenomenology of the proton-lead collisions performed in winter 2013 at the LHC and whose data are about to be available. I compute the di-gluon production cross-section which provides the simplest observable for finding quantitative evidences of saturation in the kinematic range of the LHC. I also discuss the limit of the strongly correlated final state at large transverse momenta and by the way, generalize parton distribution to dense regime. The second main topic is the quantum evolution of the quark and gluon spectra in nucleus-nucleus collisions having in mind the proof of its universal character. This result is already known for gluons and here I detail the calculation carefully. For quarks universality has not been proved yet but I derive an intermediate leading order to next-to leading order recursion relation which is a crucial step for extracting the quantum evolution. Finally I briefly present an independent work in group theory. I detail a method I used for computing traces involving an arbitrary number of group generators, a situation often encountered in QCD calculations.