In this thesis the chromatographic behavior of acetylated peptides and peptides modified with one of the most popular labels used in quantitative proteomics (TMT or Tandem Mass Tags) is analysed in several chromatographic separation modes with the objective of developing retention prediction models. Acetylation has been selected to complement TMT study as both of these modifications alter primary amine moieties but differ in charge modification effect: TMT relocates the charge away from the peptide backbone, while acetylation removes it. Carefully designed experimental procedures allowed to generate extremely large datasets of pairs of peptides (modified with their non modified counterparts) that permitted to discover some of the composition and sequence specific features (peptide length, ion pairing effect and propensity to form a helical structure) that influence how the analytes interact with their surrounding chromatographic environment. All these features were used to develop sequence specific, position dependent, highly accurate prediction models that showed very good correlations between the predicted and experimental values (R2 correlation values varied from 0.934 to 0.989 for different separation modes). These findings are of an extreme importance, especially for the TMT modification, as it constitutes the first retention time prediction algorithm for the most popular labeling tag in quantitative proteomics.