Alternative expression analysis by RNA sequencing

Griffith, Malachi; Mwenifumbo, Jill; Goya, Rodrigo; Morrissy, A. Sorana; Morin, Ryan D.; Corbett, Richard; Tang, Michelle; Hou, Ying-Chen Claire; Pugh, Trevor J.; Robertson, Gordon; Chittaranjan, Suganthi; Ally, Adrian; Asano, Jennifer; Chan, Susanna; Li, Haiyan I.; McDonald, Helen; Teague, Kevin; Zhao, Yongjun; Zeng, Thomas; Delaney, Allen; Hirst, Martin; Morin, Gregg B.; Jones, Steven J.M.; Tai, Isabella T.; Marra, Marco A.

doi:10.1038/nmeth.1503

Cited by 275 publications

(227 citation statements)

References 30 publications

Supporting

Mentioning

213

Contrasting

Unclassified

Order By: Relevance

“…Up until now, much software is available for gene expression analysis based on the RNA-Seq data (Table 3). Some is designed for quantifying the expression of known genes or isoforms and some others do not need the prior gene structure annotation information [7,10,[35][36][37][38][39]. Cufflinks [7] assembles the alignments into a parsimonious set of transcripts and then estimates the relative abundances of these transcripts based on how many reads are mapped onto them.…”

Section: Gene and Isoform Expression Quantificationmentioning

confidence: 99%

“…MISO (Mixture of Isoforms) [36] is a probabilistic framework and uses the inferred assignment of reads to isoforms to estimate the abundances of those isoforms. ALEXA-Seq [35] is a method for alternative expression analysis and also can quantify the expression of isoforms. Besides these algorithms, there are also other software that can be used for the gene expression analysis (Table 3).…”

Section: Gene and Isoform Expression Quantificationmentioning

confidence: 99%

See 1 more Smart Citation

Overview of available methods for diverse RNA-Seq data analyses

Chen

Wang

Shi

2011

Sci. China Life Sci.

View full text Add to dashboard Cite

RNA-Seq technology is becoming widely used in various transcriptomics studies; however, analyzing and interpreting the RNA-Seq data face serious challenges. With the development of high-throughput sequencing technologies, the sequencing cost is dropping dramatically with the sequencing output increasing sharply. However, the sequencing reads are still short in length and contain various sequencing errors. Moreover, the intricate transcriptome is always more complicated than we expect. These challenges proffer the urgent need of efficient bioinformatics algorithms to effectively handle the large amount of transcriptome sequencing data and carry out diverse related studies. This review summarizes a number of frequently-used applications of transcriptome sequencing and their related analyzing strategies, including short read mapping, exon-exon splice junction detection, gene or isoform expression quantification, differential expression analysis and transcriptome reconstruction. To date, RNA-Seq has been applied to a number of species for various research, such as inferring alternative splicing [4,5], quantifying the expression of genes and transcripts [6,7], detecting gene fusions [8,9], revealing long noncoding RNAs (lncRNAs) [10], and identifying single nucleotide variants (SNVs) in expressed exons [11]. Although RNA-Seq has brought tremendous benefits to these studies, it also faces many challenges from both sequencing technologies themselves and bioinformatics analyses of the data. In detail, RNA-Seq has biases in library construction, and strand-specific libraries are still not easy to be produced but are important for determining the orientation of transcripts [1]. Furthermore, RNA-Seq generates a large amount of data, and the read length is generally short and sequencing errors exist in the reads. These aspects challenge the corresponding methods and algorithms to effectively process the large amount of RNA-Seq data.

show abstract

Section: Gene and Isoform Expression Quantificationmentioning

confidence: 99%

Section: Gene and Isoform Expression Quantificationmentioning

confidence: 99%

Overview of available methods for diverse RNA-Seq data analyses

Chen

Wang

Shi

2011

Sci. China Life Sci.

View full text Add to dashboard Cite

show abstract

“…For case-control studies, several differential transcript expression (DTE) analysis methods such as Cuffdiff 2 (Trapnell et al, 2013), DESeq2 (Love et al, 2014), edgeR (Robinson et al, 2010a) and ALEXA-Seq (Griffith et al, 2010) Original Paper to discover genes that have differentially expressed transcripts whose abundance values alter between known biological conditions. In addition to the DTE methods, differential splicing (DS) analysis methods such as MISO (Katz et al, 2010), FDM (Singh et al, 5 2011), MATS (Shen et al, 2012), DEXSeq (Anders et al, 2012) and DiffSplice (Hu et al, 2013) are focused on identifying difference in relative abundance of transcripts.…”

Section: Introductionmentioning

confidence: 99%

SDEAP: a splice graph based differential transcript expression analysis tool for population data

Yang

Jiang

2016

Bioinformatics

View full text Add to dashboard Cite

Motivation: Differential transcript expression (DTE) analysis without predefined conditions is critical to biological studies. For example, it can be used to discover biomarkers to classify cancer samples into previously unknown subtypes such that better diagnosis and therapy methods can be developed for the subtypes. Although several DTE tools for population data, i.e. data without 20 known biological conditions, have been published, these tools either assume binary conditions in the input population or require the number of conditions as a part of the input. Fixing the number of conditions to binary is unrealistic and may distort the results of a DTE analysis. Estimating the correct number of conditions in a population could also be challenging for a routine user. Moreover, the existing tools only provide differential usages of exons, which may be insufficient to 25 interpret the patterns of alternative splicing across samples and restrains the applications of the tools from many biology studies.Results: We propose a novel DTE analysis algorithm, called SDEAP, that estimates the number of conditions directly from the input samples using a Dirichlet mixture model and discovers alternative splicing events using a new graph modular decomposition algorithm. By taking advantage of the above 30 technical improvement, SDEAP was able to outperform the other DTE analysis methods in our extensive experiments on simulated data and real data with qPCR validation. The prediction of SDEAP also allowed us to classify the samples of cancer subtypes and cell-cycle phases more accurately.

show abstract

“…Although this is fundamentally true, as shown in studies on how RNA-seq data corresponds with microarray and RT-PCR data [18,53,92], there is still some work to be done to fully understand the characteristics of RNA-seq data and to properly process them in order to obtain accurate results in further statistical analysis.…”

Section: Discussionmentioning

confidence: 99%

“…Regarding the RT-PCR measurements from these two experiments, we identified positive (RT-PCR differentially expressed) and negative (RT-PCR non-differentially expressed) genes following a previously reported procedure [18,53] and matched them to Ensembl IDs. After discarding replicates and eliminating unmatched genes, a total of 330 and 82 positive genes and 83 and 12 negative genes for MAQC and Griffith's dataset, respectively, were taken to compute the indicators for the performance plots.…”

Section: Differential Expression In Rna-seqmentioning

confidence: 99%

Statistical methods for transcriptomics: From microarrays to RNA-seq

Tarazona¹

View full text Add to dashboard Cite

One of the most common types of analysis in genome research is the comparison of gene expression profiles (or transcriptomics) to understand the relationship between genes (or genotype) and the phenotype. Transcriptome analysis has been traditionally conducted using microarrays and with increasing frequency since 2008 by RNA sequencing (RNA-seq). A fundamental goal in these types of studies is to identify the genes whose expression changes between different conditions, in other words, to select the most relevant variables (genes) in terms of inter-condition variability.The variable selection problem, usually known in transcriptomics as "differential expression analysis", can be addressed from the univariate or multivariate point of view, but must always take the complexity of the experimental design into account. One of the challenges that biostatisticians face when tackling this problem is the so-called "curse of dimensionality": hundreds or even thousands of variables have to be analyzed with few observations usually available. Therefore, it is essential to provide researchers with efficient statistical tools to perform this task.A typical approach to variable selection is to test the null hypothesis of equality of average expression levels between two conditions in a gene-wise fashion and to do this repeatedly for all genes in the transcriptomics data set. However, transcriptomics may also involve multifactorial experimental designs (eg multiple treatments, several developmental states, time series...). The first part of this thesis is dedicated to the variable selection problem in multifactorial designs when multivariate methods are used to model microarray gene expression profiles. In particular, we chose the ASCA-genes multivariate technique as a starting point to propose some strategies to select differentially expressed genes in the multivariate context, that were tested under different biological scenarios.RNA-seq technology emerged when work for this thesis was first started and now it is widely applied in transcriptomics. RNA-seq data is fundamentally different from microarray data and this has motivated the development of new statistical methods to study differential expression in this technology. Thus, the second part of the thesis is entirely focused on RNA-seq experiments. First, we develop a set of procedures to assess the quality of RNA-seq measurements, to identify the potential biases of the technology and to process the data to reduce the impact of technical noise on statistical results. Secondly, we address the variable selection problem for the two-class comparison case. Given that some controversy exists on the theoretical distribution followed by RNA-seq data, we opted to investigate non-parametric datadriven techniques to overcome the limitations of parametric assumptions and propose a strategy that is efficient in controlling the false positive rate. Two methodologies, NOISeq for technical and NOISeqBIO for biological replicates, were developed and compared to the state of th...

show abstract

Alternative expression analysis by RNA sequencing

Cited by 275 publications

References 30 publications

Overview of available methods for diverse RNA-Seq data analyses

Overview of available methods for diverse RNA-Seq data analyses

SDEAP: a splice graph based differential transcript expression analysis tool for population data

Statistical methods for transcriptomics: From microarrays to RNA-seq

Contact Info

Product

Resources

About