Genome Research · 2003 · 734 citations · 32 references
Life SciencesBioinformaticsBiostatisticsGene ExpressionGene Expression AnalysisGenomic Data AnalysisCancer GenomicsCancer TranscriptomicsGene Expression ProfilingMolecular Tumor ProfilingTranscriptomic ProfilingCancer BioinformaticsComputational Linear AlgebraConditionsGene ClusteringGene Expression SignaturesMarker Gene ExpressionMicroarray AnalysisSingular Value Decomposition
RNA expression data are routinely used to classify genes and phenotypes, and identifying marker genes that are differentially expressed across specific conditions—often forming checkerboard patterns—has become especially relevant for distinguishing tumor subtypes. The authors aim to develop a biclustering method that simultaneously clusters genes and conditions to uncover checkerboard patterns of differential expression. Spectral biclustering identifies checkerboard structures by extracting characteristic eigenvectors via singular value decomposition, with variants that normalize genes and conditions independently or jointly, and the authors evaluate these methods on public cancer datasets against benchmarks such as raw SVD and normalized cuts.
Global analyses of RNA expression levels are useful for classifying genes and overall phenotypes. Often these classification problems are linked, and one wants to find “marker genes” that are differentially expressed in particular sets of “conditions.” We have developed a method that simultaneously clusters genes and conditions, finding distinctive “checkerboard” patterns in matrices of gene expression data, if they exist. In a cancer context, these checkerboards correspond to genes that are markedly up- or downregulated in patients with particular types of tumors. Our method, spectral biclustering, is based on the observation that checkerboard structures in matrices of expression data can be found in eigenvectors corresponding to characteristic expression patterns across genes or conditions. In addition, these eigenvectors can be readily identified by commonly used linear algebra approaches, in particular the singular value decomposition (SVD), coupled with closely integrated normalization steps. We present a number of variants of the approach, depending on whether the normalization over genes and conditions is done independently or in a coupled fashion. We then apply spectral biclustering to a selection of publicly available cancer expression data sets, and examine the degree to which the approach is able to identify checkerboard structures. Furthermore, we compare the performance of our biclustering methods against a number of reasonable benchmarks (e.g., direct application of SVD or normalized cuts to raw data).
32
Todd R. Golub, Donna K. Slonim, Pablo Tamayo et al. · Science · 1999 · 11.6K citations
Distinct types of diffuse large B-cell lymphoma identified by gene expression profiling
Ash A. Alizadeh, Michael B. Eisen, R. Eric Davis et al. · Nature · 2000 · 10K citations
Gene expression profiling predicts clinical outcome of breast cancer
Laura van ‘t Veer, Hongyue Dai, Marc J. van de Vijver et al. · Nature · 2002 · 9.6K citations · Full text