Protein Arginine (R) methylation is a post-translational modification involved in various biological processes, such as RNA splicing, DNA repair, immune response, signal transduction, and tumour development. Although several advancements were made in the study of this modification by mass spectrometry, researchers still face the problem of a high false discovery rate. We present a dataset of high-quality methylations obtained from several different heavy methyl SILAC (hmSILAC) experiments analysed with a machine learning-based tool doublets and show that this model allows for improved high-confidence identification of real methyl-peptides. Overall, our results are consistent with the notion that protein R methylation modulates protein:RNA interactions and suggest a role in rewiring protein:protein interactions, for which we provide experimental evidence for a representative case (i.e. NONO:PSPC1). Upon intersecting our R-methyl-sites dataset with a phosphosites dataset, we observed that R methylation correlates differently with S/T-Y phosphorylation in response to various stimuli. Finally, we explored the application of hmSILAC to identify unconventional methylated residues and successfully identified novel histone methylation marks on Serine 28 and Threonine 32 of H3.