Dataset Information

Predictive response-relevant clustering of expression data provides insights into disease processes.

ABSTRACT: This article describes and illustrates a novel method of microarray data analysis that couples model-based clustering and binary classification to form clusters of `response-relevant' genes; that is, genes that are informative when discriminating between the different values of the response. Predictions are subsequently made using an appropriate statistical summary of each gene cluster, which we call the `meta-covariate' representation of the cluster, in a probit regression model. We first illustrate this method by analysing a leukaemia expression dataset, before focusing closely on the meta-covariate analysis of a renal gene expression dataset in a rat model of salt-sensitive hypertension. We explore the biological insights provided by our analysis of these data. In particular, we identify a highly influential cluster of 13 genes--including three transcription factors (Arntl, Bhlhe41 and Npas2)-that is implicated as being protective against hypertension in response to increased dietary sodium. Functional and canonical pathway analysis of this cluster using Ingenuity Pathway Analysis implicated transcriptional activation and circadian rhythm signalling, respectively. Although we illustrate our method using only expression data, the method is applicable to any high-dimensional datasets. Expression data are available at ArrayExpress (accession number E-MEXP-2514) and code is available at http://www.dcs.gla.ac.uk/inference/metacovariateanalysis/.

SUBMITTER: Hopcroft LE

PROVIDER: S-EPMC2978340 | biostudies-literature | 2010 Nov

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Predictive response-relevant clustering of expression data provides insights into disease processes.

Hopcroft Lisa E M LE McBride Martin W MW Harris Keith J KJ Sampson Amanda K AK McClure John D JD Graham Delyth D Young Graham G Holyoake Tessa L TL Girolami Mark A MA Dominiczak Anna F AF

Nucleic acids research 20100622 20

This article describes and illustrates a novel method of microarray data analysis that couples model-based clustering and binary classification to form clusters of `response-relevant' genes; that is, genes that are informative when discriminating between the different values of the response. Predictions are subsequently made using an appropriate statistical summary of each gene cluster, which we call the `meta-covariate' representation of the cluster, in a probit regression model. We first illus ...[more]

PMID: 20571087

Similar Datasets

Project description:Background – Hypertrophic cardiomyopathy (HCM) is a common genetic cardiac disorder that is characterized by the presence of left ventricular hypertrophy with histological evidence of myocyte disarray and fibrosis. Pathologic DNA variants in sarcomeric genes are seen in 50–60% of families, but the mechanistic origins of hypertrophy and disarray remain poorly understood. More molecular insights could contribute to a better understanding of these origins and improve therapeutic strategies for HCM. Methods – To identify gene expression patterns relevant to HCM pathology, we quantified intercellular heterogeneity in cardiomyocyte gene expression by performing single-cell RNA sequencing (scRNA-seq) on myectomy tissue from HCM patients. Results – Septal myectomy samples provide a unique opportunity to access diseased human heart tissue for scRNA-seq. We analysed gene expression of individual cardiomyocytes for three HCM patients, who each showed a similar presence of cardiomyocyte subpopulations. This similarity indicated these subpopulations emerge from biological processes, and not from experimental variability. Moreover, our data revealed functional links between genes important to HCM. A subpopulation characterized by elevated titin (TTN) expression showed increased expression of genes involved in muscle contraction and development. A subpopulation of Natriuretic Peptide A (NPPA)-enriched cells showed elevated expression of Natriuretic Peptide B (NPPB). Further analysis of gene co-expression revealed five groups of genes with tightly correlated expression (regulons), one consisting of mostly sarcomeric genes and another containing some well-known stress markers. These correlations were confirmed in 97 additional HCM myectomy samples and allowed us to de novo identify transcription factors related to sarcomeric gene expression and stress. Finally, using forward scatter, we charted functional implications of regulon expression and showed a link between the stress-marker regulon and cell size. Histology samples independently confirmed that regulon member skeletal alpha actin (ACTA1) correlates with cell size. Conclusion – We have shown that single-cell sequencing on myectomy samples from HCM patients identified new functional links between genes, transcription factors, and a functional link between genes and cell size. This highlights the relevance of cardiomyocyte heterogeneity, and additionally provides a wealth of novel insights into molecular events at the base of HCM, which potentially contribute to understanding disease progression.

Dataset Information

Predictive response-relevant clustering of expression data provides insights into disease processes.

Publications

Predictive response-relevant clustering of expression data provides insights into disease processes.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets