The article by Orly Alter, Patrick O. Brown, and David Botstein describes the application of singular value decomposition (SVD) in processing and modeling genome-wide expression data. SVD transforms the data from a genes × arrays space to a reduced "eigengenes" × "eigenarrays" space, where eigengenes and eigenarrays are unique orthonormal superpositions of genes and arrays, respectively. This transformation diagonalizes the data, allowing for meaningful comparison of gene expression across different arrays and experiments by filtering out noise and artifacts. The eigengenes and eigenarrays provide a global picture of gene expression dynamics, classifying genes and arrays into groups based on similar regulation, function, cellular state, or biological phenotype. The authors demonstrate the utility of SVD through examples, including the analysis of elutriation-synchronized cell cycle data and α-factor-synchronized cell cycle data with overactivated CLB2 and CLN3 genes. They show that SVD can effectively capture significant eigengenes and eigenarrays, which can be associated with observed genome-wide effects of regulators or measured samples. The method provides a robust framework for processing and modeling genome-wide expression data, making it easier to interpret biological information from large-scale expression datasets.The article by Orly Alter, Patrick O. Brown, and David Botstein describes the application of singular value decomposition (SVD) in processing and modeling genome-wide expression data. SVD transforms the data from a genes × arrays space to a reduced "eigengenes" × "eigenarrays" space, where eigengenes and eigenarrays are unique orthonormal superpositions of genes and arrays, respectively. This transformation diagonalizes the data, allowing for meaningful comparison of gene expression across different arrays and experiments by filtering out noise and artifacts. The eigengenes and eigenarrays provide a global picture of gene expression dynamics, classifying genes and arrays into groups based on similar regulation, function, cellular state, or biological phenotype. The authors demonstrate the utility of SVD through examples, including the analysis of elutriation-synchronized cell cycle data and α-factor-synchronized cell cycle data with overactivated CLB2 and CLN3 genes. They show that SVD can effectively capture significant eigengenes and eigenarrays, which can be associated with observed genome-wide effects of regulators or measured samples. The method provides a robust framework for processing and modeling genome-wide expression data, making it easier to interpret biological information from large-scale expression datasets.