Comparative evaluation of isoform-level gene expression estimation algorithms for RNA-seq and exon-array platforms

Abstract

Given that the majority of multi-exon genes generate diverse functional products, it is important to evaluate expression at the isoform level. Previous studies have demonstrated strong gene-level correlations between RNA sequencing (RNA-seq) and microarray platforms, but have not studied their concordance at the isoform level. We performed transcript abundance estimation on raw RNA-seq and exon-array expression profiles available for common glioblastoma multiforme samples from The Cancer Genome Atlas using different analysis pipelines, and compared both the isoform- and gene-level expression estimates between programs and platforms. The results showed better concordance between RNA-seq/exon-array and reverse transcription-quantitative polymerase chain reaction (RT-qPCR) platforms for fold change estimates than for raw abundance estimates, suggesting that fold change normalization against a control is an important step for integrating expression data across platforms. Based on RT-qPCR validations, eXpress and Multi-Mapping Bayesian Gene eXpression (MMBGX) programs achieved the best performance for RNA-seq and exon-array platforms, respectively, for deriving the isoform-level fold change values. While eXpress achieved the highest correlation with the RT-qPCR and exon-array (MMBGX) results overall, RSEM was more highly correlated with MMBGX for the subset of transcripts that are highly variable across the samples. eXpress appears to be most successful in discriminating lowly expressed transcripts, but IsoformEx and RSEM correlate more strongly with MMBGX for highly expressed transcripts. The results also reinforce how potentially important isoform-level expression changes can be masked by gene-level estimates, and demonstrate that exon arrays yield comparable results to RNA-seq for evaluating isoform-level expression changes.

The Transcription Factor Nfatc2 Regulates β-Cell Proliferation and Genes Associated with Type 2 Diabetes in Mouse and Human Islets

“…Isoforms were quantified by running RSEM (version 1.2.4) [65] and Bowtie1 (version 0.12.7) [66] in paired-end mode and using a synthetically reconstructed transcriptome derived from the mm9 reference genome and RefSeq gene models (downloaded from UCSC browser June 24 th , 2013, NR entries removed). The Bowtie-RSEM pipeline directly maps RNA reads to annotated transcripts (isoforms), which has been show to provide better quantification accuracy for known transcripts compared to pipelines that uses splice aligner to map reads to the whole genome [65, 69]. All other RSEM parameters used were default.…”

Systematic Identification of Housekeeping Genes Possibly Used as References in Caenorhabditis elegans by Large-Scale Data Integration

“…With the development of high throughput expression technologies (microarray and RNA-seq), massive gene expression datasets are produced daily. There is a good agreement between RNA-seq and microarray relative to gene expression, despite some data variability in low-expression genes that may be due to the difference in expression platform and data analysis [61,62]. Microarray data were used in this study by considering the following reasons: (1) microarray is a mature genomic platform with a well-established data analysis pipeline; (2) microarray is reliable in model organisms [63] and we just focus on well-annotated genes in model organism C. elegans; and (3) despite the fact that RNA-seq enables to identify non-coding differentially expressed genes or not well-annotated genes that offer a potential for improved mechanistic clarity, we do not consider those genes.…”

The long non-coding RNAGHSROSreprograms prostate cancer cell lines toward a more aggressive phenotype

“…Exon arrays are comparable to RNA-seq in experiments aimed at assessing exon expression (i.e. gene isoforms) and suitable for experiments where the exon of interest is known 121,122 . In the Exon 1.0 ST array, known (genes and ESTs) and putative exons are combined to form 'transcript clusters', with each exon defined as a probe set (typically, a set of 2-4 probes).…”