Discovery-based proteomics can be thought of as an attempt to understand the information encoded in genomic sequences from the perspective of proteins (i.e. the structural, functional and regulatory mechanisms imparted to a cell by the expressed proteins). In practice, it contrasts with the hypothesis-driven, serial approach that was successful for protein chemists in the last century. The technology today is built on the basic understanding of the physicochemical properties of proteins developed during that era. In addition, proteomics attempts to study biological processes comprehensively or globally by systematic, parallel analysis of many properties associated with proteins expressed in a cell. While there are many analytical techniques in use and under development in proteomics, mass spectrometry (MS) is currently one of the field's most important discovery-based tools. For the most part, mass spectrometers have been used to catalog proteins, but recently the role of MS in proteomics was extended beyond descriptive analysis to include measurement of differences in protein expression between cellular states. These experiments involve comparison of proteins labeled in vitro or in vivo with either "heavy" or "light" stable isotopes to provide internal standards for relative quantification. Together with MS, differential stable-isotope labeling of peptides allows an experiment to be conducted for proteins to provide information on the differences in protein expression between cells in different environments. While the experiments are quite different, their intent is similar to using cDNA arrays to measure changes in mRNA expression (i.e. what are the global changes proteins concomitant with the observable or known phenotypic definition of a given cellular state relative to another?) This article will review the use of MS and stable isotopic labeling of peptides to detect differences in protein expression between two cellular states. (C) 2003 Published by Elsevier Science B.V.