Fischer-Vize, Science 270, 1828 (1995).35. T. C. James and S. C. Elgin, Mol. Cell Biol. 6, 3862(1986); R. Paro and D. S. Hogness, Proc. Natl. Acad.Sci. U.S.A. 88, 263 (1991); B. Tschiersch et al.,EMBO J. 13, 3822 (1994); M. T. Madireddi et al., Cell87, 75 (1996); D. G. Stokes, K. D. Tartof, R. P. Perry,Proc. Natl. Acad. Sci. U.S.A. 93, 7137 (1996).36. P. M. Palosaari et al., J. Biol. Chem. 266, 10750(1991); A. Schmitz, K. H. Gartemann, J. Fiedler, E.Grund, R. Eichenlaub, Appl. Environ. Microbiol. 58,4068 (1992); V. Sharma, K. Suvarna, R. Mega-nathan, M. E. Hudspeth, J. Bacteriol. 174, 5057(1992); M. Kanazawa et al., Enzyme Protein 47,9(1993); Z. L. Boynton, G. N. Bennet, F. B. Rudolph,J. Bacteriol. 178, 3015 (1996).37. M. Ho et al., Cell 77, 869 (1994).38. W. Hendriks et al., J. Cell Biochem. 59, 418 (1995).39. We thank H. Skaletsky and F. Lewitter for help withsequence analysis; Lawrence Livermore NationalLaboratory for the flow-sorted Y cosmid library; andP. Bain, A. Bortvin, A. de la Chapelle, G. Fink, K.Jegalian, T. Kawaguchi, E. Lander, H. Lodish, P.Matsudaira, D. Menke, U. RajBhandary, R. Reijo, S.Rozen, A. Schwartz, C. Sun, and C. Tilford for com-ments on the manuscript. Supported by NIH.28 April 1997; accepted 9 September 1997Exploring the Metabolic and Genetic Control ofGene Expression on a Genomic ScaleJoseph L. DeRisi, Vishwanath R. Iyer, Patrick O. Brown*DNA microarrays containing virtually every gene of Saccharomyces cerevisiae were usedto carry out a comprehensive investigation of the temporal program of gene expressionaccompanying the metabolic shift from fermentation to respiration. The expressionprofiles observed for genes with known metabolic functions pointed to features of themetabolic reprogramming that occur during the diauxic shift, and the expression patternsof many previously uncharacterized genes provided clues to their possible functions. Thesame DNA microarrays were also used to identify genes whose expression was affectedby deletion of the transcriptional co-repressor TUP1 or overexpression of the transcrip-tional activator YAP1. These results demonstrate the feasibility and utility of this ap-proach to genomewide exploration of gene expression patterns.The complete sequences of nearly a dozenmicrobial genomes are known, and in thenext several years we expect to know thecomplete genome sequences of severalmetazoans, including the human genome.Defining the role of each gene in thesegenomes will be a formidable task, and un-derstanding how the genome functions as awhole in the complex natural history of aliving organism presents an even greaterchallenge.Knowing when and where a gene isexpressed often provides a strong clue as toits biological role. Conversely, the patternof genes expressed in a cell can providedetailed information about its state. Al-though regulation of protein abundance ina cell is by no means accomplished solelyby regulation of mRNA, virtually all dif-ferences in cell type or state are correlatedwith changes in the mRNA levels of manygenes. This is fortuitous because the onlyspecific reagent required to measure theabundance of the mRNA for a specificgene is a cDNA sequence. DNA microar-rays, consisting of thousands of individualgene sequences printed in a high-densityarray on a glass microscope slide (1, 2),provide a practical and economical toolfor studying gene expression on a verylarge scale (3–6).Saccharomyces cerevisiae is an especiallyfavorable organism in which to conduct asystematic investigation of gene expression.The genes are easy to recognize in the ge-nome sequence, cis regulatory elements aregenerally compact and close to the tran-scription units, much is already knownabout its genetic regulatory mechanisms,and a powerful set of tools is available for itsanalysis.A recurring cycle in the natural historyof yeast involves a shift from anaerobic(fermentation) to aerobic (respiration) me-tabolism. Inoculation of yeast into a medi-um rich in sugar is followed by rapid growthfueled by fermentation, with the productionof ethanol. When the fermentable sugar isexhausted, the yeast cells turn to ethanol asa carbon source for aerobic growth. Thisswitch from anaerobic growth to aerobicrespiration upon depletion of glucose, re-ferred to as the diauxic shift, is correlatedwith widespread changes in the expressionof genes involved in fundamental cellularprocesses such as carbon metabolism, pro-tein synthesis, and carbohydrate storage(7). We used DNA microarrays to charac-terize the changes in gene expression thattake place during this process for nearly theentire genome, and to investigate the ge-netic circuitry that regulates and executesthis program.Yeast open reading frames (ORFs) wereamplified by the polymerase chain reaction(PCR), with a commercially available set ofprimer pairs (8). DNA microarrays, con-taining approximately 6400 distinct DNAsequences, were printed onto glass slides byusing a simple robotic printing device (9).Cells from an exponentially growing cultureof yeast were inoculated into fresh mediumand grown at 30°C for 21 hours. After aninitial 9 hours of growth, samples were har-vested at seven successive 2-hour intervals,and mRNA was isolated (10). Fluorescentlylabeled cDNA was prepared by reverse tran-scription in the presence of Cy3(green)-or Cy5(red)-labeled deoxyuridine triphos-phate (dUTP) (11) and then hybridized tothe microarrays (12). To maximize the re-liability with which changes in expressionlevels could be discerned, we labeled cDNAprepared from cells at each successive timepoint with Cy5, then mixed it with a Cy3-labeled “reference” cDNA sample preparedfrom cells harvested at the first intervalafter inoculation. In this experimental de-sign, the relative fluorescence intensitymeasured for the Cy3 and Cy5 fluors ateach array element provides a reliable mea-sure of the relative abundance of the corre-sponding mRNA in the two cell popula-tions (Fig. 1). Data from the series of sevensamples (Fig. 2), consisting of more than43,000 expression-ratio measurements,were organized into a database to facilitateefficient exploration and analysis of theresults. This database is publicly availableon the Internet (13).During exponential growth in glucose-rich medium, the global pattern of geneexpression was remarkably stable. Indeed,when gene expression patterns between thefirst two cell samples (harvested at a 2-hourinterval) were compared, mRNA levels dif-fered by a factor of 2 or more for only 19genes