Mixture modeling of microarray gene expression data

Abstract

About 28% of genes appear to have an expression pattern that follows a mixture distribution. We use first- and second-order partial correlation coefficients to identify trios and quartets of non-sex-linked genes that are highly associated and that are also mixtures. We identified 18 trio and 35 quartet mixtures and evaluated their mixture distribution concordance. Concordance was defined as the proportion of observations that simultaneously fall in the component with the higher mean or simultaneously in the component with the lower mean based on their Bayesian posterior probabilities. These trios and quartets have a concordance rate greater than 80%. There are 33 genes involved in these trios and quartets. A factor analysis with varimax rotation identifies three gene groups based on their factor loadings. One group of 18 genes has a concordance rate of 56.7%, another group of 8 genes has a concordance rate of 60.8%, and a third group of 7 genes has a concordance rate of 69.6%. Each of these rates is highly significant, suggesting that there may be strong biological underpinnings for the mixture mechanisms of these genes. Bayesian factor screening confirms this hypothesis by identifying six single-nucleotide polymorphisms that are significantly associated with the expression phenotypes of the five most concordant genes in the first group.

Bimodal distribution of RNA expression levels in human skeletal muscle tissue

“…Our estimate also differs greatly from the only other prevalence reported for bimodal expression in a population comprised solely of healthy persons - 28% in lymphoblastoid cells [18,28]. As the threshold for assessing bimodality in that study was much more liberal, we re-analyzed the data from this population and found only 2.9% of genes of the limited gene set previously analyzed to show bimodality consistent with the thresholds used in our analysis (see Methods).…”

“…Limited reports of bimodal expression patterns in other studies have also been made [ 18 - 20 ], particularly when heterogeneous tissues are being contrasted [ 21 - 26 ] such as between healthy and diseased tissue (and for which bimodal expression levels would be expected for perhaps many genes). However, in homogeneous tissue of apparently healthy individuals, such bimodality may be surprising, and hints at unrecognized heterogeneity which may involve precipitating factors of future disease development.…”

Data mining of RNA expression and DNA genotype data: Presentation Group 5 contributions to Genetic Analysis Workshop 15

“…Gong et al [2007] reported a number of hub SNPs and hub genes. Yang et al [2007] reported six SNPs that were simultaneous predictors of gene expression levels in genes AHSA1, ELAC2, CCT3, TUBG1, and TACC3 as well as evidence of a common mixture mechanism simultaneously affecting a number of genes.…”

“…For example, Liu et al [2007] focused on two genes, one new and one from a previous paper; Yang et al [2007] selected distributions with mixtures; and Zhao et al [2007] selected multivariate normals. Given this diversity and the richness of the dataset, one might expect that the findings would have little overlap, much as the blind men of Indostan describing an elephant, where each focused on a different aspect of the elephant.…”

“…McLachlan and coworkers have proposed using mixture modeling to analyze microarray data [McLachlan et al, 2002[McLachlan et al, , 2003. Yang et al [2007] used this approach and found that about 28% of gene expression variables appeared to be a mixture of two or more components using the SKUMIX algorithm [Maclean et al, 1976]. They identified sets of three or four genes involving 33 autosomal genes that appeared to have a mixture distribution and met partial correlation coefficient criteria.…”