A flexible count data model to fit the wide diversity of expression profiles arising from extensively replicated RNA-seq experiments

Abstract

BackgroundHigh-throughput RNA sequencing (RNA-seq) offers unprecedented power to capture the real dynamics of gene expression. Experimental designs with extensive biological replication present a unique opportunity to exploit this feature and distinguish expression profiles with higher resolution. RNA-seq data analysis methods so far have been mostly applied to data sets with few replicates and their default settings try to provide the best performance under this constraint. These methods are based on two well-known count data distributions: the Poisson and the negative binomial. The way to properly calibrate them with large RNA-seq data sets is not trivial for the non-expert bioinformatics user.ResultsHere we show that expression profiles produced by extensively-replicated RNA-seq experiments lead to a rich diversity of count data distributions beyond the Poisson and the negative binomial, such as Poisson-Inverse Gaussian or Pólya-Aeppli, which can be captured by a more general family of count data distributions called the Poisson-Tweedie. The flexibility of the Poisson-Tweedie family enables a direct fitting of emerging features of large expression profiles, such as heavy-tails or zero-inflation, without the need to alter a single configuration parameter. We provide a software package for R called implementing a new test for differential expression based on the Poisson-Tweedie family. Using simulations on synthetic and real RNA-seq data we show that yields P-values that are equally or more accurate than competing methods under different configuration parameters. By surveying the tiny fraction of sex-specific gene expression changes in human lymphoblastoid cell lines, we also show that accurately detects differentially expressed genes in a real large RNA-seq data set with improved performance and reproducibility over the previously compared methodologies. Finally, we compared the results with those obtained from microarrays in order to check for reproducibility.ConclusionsRNA-seq data with many replicates leads to a handful of count data distributions which can be accurately estimated with the statistical model illustrated in this paper. This method provides a better fit to the underlying biological variability; this may be critical when comparing groups of RNA-seq samples with markedly different count data distributions. The package forms part of the Bioconductor project and it is available for download at http://www.bioconductor.org.

RNA-Seq Technology and Its Application in Fish Transcriptomics

“…The RNA-seq data corroborated most of the microarray results and detected novel transcripts that were not captured by microarray. This discrepancy in results between microarray and RNA-seq was also observed in other researches and might be improved by statistical methods with proper assessment the statistical significance of the observed changes (Esnaola et al, 2013). In another study, the toxicological effects of perfluorooctane sulfonate (PFOS), a widely-distributed persistent organic pollutant, on Oryzias melastigma embryos were examined using RNA-seq (Huang et al, 2012).…”

“…There is no standard method to detect DE due to the short history of RNA-seq technology. The currently popular tools for DE analysis in Bioconductor include edger , DESeq (Anders and Huber, 2010), baySeq (Hardcastle and Kelly, 2010), and tweeDEseq (Esnaola et al, 2013), and methods, not included in Bioconductor, also exist, such as ShrinkBayes (Van De Wiel et al, 2013) and TSPM (Auer and Doerge, 2011). These DE analysis tools have already been reviewed in detail and well compared in their gene ranking performances (Dillies et al, 2012;Kvam et al, 2012;Soneson et al, 2013).…”

Error estimates for the analysis of differential expression from RNA-seq count data

“…Accordingly we show in Figure 6 analogous simulations using Poisson-inverse-Gaussian (PIG) data, which is one of the class of Poisson-Tweedie distributions often used to simulate overdispersed integer-count data. There is evidence from RNA-seq data with very large numbers of replicates that the PIG distribution may be at least as good a fit as NB for many of genes within the transcriptome (Esnaola et al, 2013). This distribution is closer to Poisson and should therefore not discriminate as much against the package PoissonSeq as the NB distribution.…”

RNA-seq 2G: online analysis of differential gene expression with comprehensive options of statistical methods

“…We evaluated and tested over 40 DE methods (4)(5)(6)(7)(8)(11)(12)(13)(14)(15)(16)(17)(18)(19)(20)(21)(22)(23)(24)(25). About one-third of them were not selected for various reasons.…”