The quantification of differences in the proteome between physiological states is a challenging technical task. Two-dimensional polyacrylamide gel electrophoresis (2-DE) is still the standard method to separate complex protein mixtures. However, the 2-DE gel approach is insufficient to analyze entire proteomes, including very low abundance proteins. Some problems are overcome by LC/MS-based methods, but mass spectrometry is not inherently quantitative because the peptides can have differences in spectrometric response. In this study, proteins were extracted from leafs and roots of Coffea canephora (Clone 14, drought-tolerant and clone 109A, drought-sensitive) unstressed and submitted to a severe water deficit. Changes in protein profiles and concentrations were analyzed in both genotypes and water regimes. Protein extracts were separated by 2-DE and proteins differentially expressed were identified using MALDI-ToF/ToF mass spectrometry (MS) of tryptic peptides. Trans-Proteomic Pipeline (TPP) was used for analyzing, validating, and storing protein identification data. The X!Tandem software was used for MS/MS ion search against a EST-based database of Coffea. Alternatively, the protein extracts were treated with trypsin and peptides were pre-fractionated by pI before analysis by 2D-LC coupled with a Q-tof mass spectrometer. The results from different experiments were processed and compared by using differential label-free quantitative analysis with the LC-MS imaging based quantification softwares msInspect (http://proteomics.fhcrc.org) and SuperHirn (http://tools.proteomecenter.org). (Texte intégral)