Bayesian model accounting for within-class biological variability in Serial Analysis of Gene Expression (SAGE)

Vêncio, Ricardo Z. N.; Brentani, Helena; Patrão, Diogo Ferreira da Costa; Pereira, Carlos Alberto

doi:10.1186/1471-2105-5-119

Cited by 55 publications

(30 citation statements)

References 22 publications

(30 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In MPSS, signature-sampling fluctuations alone cannot account for the involved biochemical manipulations taking place in the process of signature library production. More elaborate Bayesian approaches represent the distribution of SAGE signature counts as mixtures of binomial (20) or Poisson (18) statistics whose parameters are weighted with prior distributions (chosen using heuristic criteria or mathematical convenience), yielding an enhanced variance when compared with plain Bernoulli or Poisson statistics. It is unlikely, however, that the basic process at play is a Bernoulli-type process, because there are many manipulations done to the signatures before they are counted that inherently contribute to its sample-to-sample count variability.…”

Section: Resultsmentioning

confidence: 99%

Statistical analysis of MPSS measurements: Application to the study of LPS-activated macrophage gene expression

Stolovitzky

Kundaje

Held

et al. 2005

Proc. Natl. Acad. Sci. U.S.A.

View full text Add to dashboard Cite

Massively Parallel Signature Sequencing (MPSS), a recently developed high-throughput transcription profiling technology, has the ability to profile almost every transcript in a sample without requiring prior knowledge of the sequence of the transcribed genes. As is the case with DNA microarrays, effective data analysis depends crucially on understanding how noise affects measurements. We analyze the sources of noise in MPSS and present a quantitative model describing the variability between replicate MPSS assays. We use this model to construct statistical hypotheses that test whether an observed change in gene expression in a pair-wise comparison is significant. This analysis is then extended to the determination of the significance of changes in expression levels measured over the course of a time series of measurements. We apply these analytic techniques to the study of a time series of MPSS gene expression measurements on LPS-stimulated macrophages. To evaluate our statistical significance metrics, we compare our results with published data on macrophage activation measured by using Affymetrix GeneChips.transcription profiling ͉ noise model

show abstract

Section: Resultsmentioning

confidence: 99%

Statistical analysis of MPSS measurements: Application to the study of LPS-activated macrophage gene expression

Stolovitzky

Kundaje

Held

et al. 2005

Proc. Natl. Acad. Sci. U.S.A.

View full text Add to dashboard Cite

show abstract

“…To evaluate the statistical significance of the differences in counts, a web-based program, SAGEbetaBin (Vêncio et al , 2004), was employed. The paired raw-count data sets (Cd0 versus Cd3h, Cd1d, or Cd3d) were subjected to the SAGEbetaBin analysis, and up- and down-regulated tag species were detected with a Bayes error rate of 0.05.…”

Section: Methodsmentioning

confidence: 99%

Gene expression analysis in cadmium-stressed roots of a low cadmium-accumulating solanaceous plant, Solanum torvum

Yamaguchi

Fukuoka

Arao

et al. 2009

Journal of Experimental Botany

100

View full text Add to dashboard Cite

Solanum torvum Sw. cv. Torubamubiga (TB) is a low cadmium (Cd)-accumulating plant. To elucidate the molecular mechanisms of the Cd acclimation process in TB roots, transcriptional regulation was analysed in response to mild Cd treatment: 0.1 μM CdCl2 in hydroponic solution. A unigene set consisting of 6296 unigene sequences was constructed from 18 816 TB cDNAs. The distribution of functional categories was similar to tomato, while 330 unigenes were suggested to be TB specific. For expression profiling, the SuperSAGE method was adapted for use with Illumina sequencing technology. Expression tag libraries were constructed from Cd-treated (for 3 h, 1 d, and 3 d) and untreated roots, and 34 269 species of independent tags were collected. Moreover, 6237 tags were ascribed to the TB or eggplant (aubergine) unigene sequences. Time-course changes were examined, and 2049 up- and 2022 down-regulated tags were identified. Although no tags annotated to metal transporter genes were significantly regulated, a tag annotated to AtFRD3, a xylem-loading citrate transporter, was down-regulated. In addition to induction of heavy metal chaperone proteins, antioxidative and sulphur-assimilating enzymes were induced, confirming that oxidative stress developed even using a mild Cd concentration. Rapid repression of dehydration-related transcription factors and aquaporin isoforms suggests that dehydration stress is a potential constituent of Cd-induced biochemical impediments. These transcriptional changes were also confirmed by real-time reverse transcription-PCR. Further additions of TB unigene sequences and functional analysis of the regulated tags will reveal the molecular basis of the Cd acclimation process, including the low Cd-accumulating characteristics of TB.

show abstract

“…In such cases, Poisson models are likely to produce high false positives. Some of the commonly used approaches such as, Bayesian methods (Baggerly et al, 2004;Vêncio et al, 2004), generalized linear models (Baggerly et al, 2003;Srivastava and Chen, 2010) including NB models (Anders and Huber, 2010;Smyth, 2007, 2008) assume that all gene counts are derived from an over-dispersed distribution, and fail to address the fact that some genes might have constant levels of transcription within treatment groups (Auer and Doerge, 2011;Oshlack et al, 2010). Accordingly, the assumptions on the variations of each gene's expression can mislead the detection of a truly DE gene.…”

Section: Introductionmentioning

confidence: 99%

Zero-Inflated Models for RNA-Seq Count Data

Mahi

Begum

2018

J. biomed. anal.

View full text Add to dashboard Cite

One of the main objectives of many biological studies is to explore differential gene expression profiles between samples. Genes are referred to as differentially expressed (DE) if the read counts change across treatments or conditions systematically. Poisson and negative binomial (NB) regressions are widely used methods for non-over-dispersed (NOD) and over-dispersed (OD) count data respectively. However, in the presence of excessive number of zeros, these methods need adjustments. In this paper, we consider a zero-inflated Poisson mixed effects model (ZIPMM) and zero-inflated negative binomial mixed effects model (ZINBMM) to address excessive zero counts in the NOD and OD RNA-seq data respectively in the presence of random effects. We apply these methods to both simulated and real RNA-seq datasets. The ZIPMM and ZINBMM perform better on both simulated and real datasets.

show abstract

Bayesian model accounting for within-class biological variability in Serial Analysis of Gene Expression (SAGE)

Cited by 55 publications

References 22 publications

Statistical analysis of MPSS measurements: Application to the study of LPS-activated macrophage gene expression

Statistical analysis of MPSS measurements: Application to the study of LPS-activated macrophage gene expression

Gene expression analysis in cadmium-stressed roots of a low cadmium-accumulating solanaceous plant, Solanum torvum

Zero-Inflated Models for RNA-Seq Count Data

Contact Info

Product

Resources

About