Removing technical variability in RNA-seq data using conditional quantile normalization

By Kasper D. Hansen, Rafael A. Irizarry, Zhijin Wu. Biostatistics, 2012

639 citations recorded by the source corpus.

The ability to measure gene expression on a genome-wide scale is one of the most promising accomplishments in molecular biology. Microarrays, the technology that first permitted this, were riddled with problems due to unwanted sources of variability. Many of these problems are now mitigated, after a decade's worth of statistical methodology development. The recently developed RNA sequencing (RNA-seq) technology has generated much excitement in part due to claims of reduced variability in comparison to microarrays. However, we show that RNA-seq data demonstrate unwanted and obscuring variability similar to what was first observed in microarrays. In particular, we find guanine-cytosine content (GC-content) has a strong sample-specific effect on gene expression measurements that, if left uncorrected, leads to false positives in downstream results. We also report on commonly observed data distortions that demonstrate the need for data normalization. Here, we describe a statistical methodology that improves precision by 42% without loss of accuracy. Our resulting conditional quantile normalization algorithm combines robust generalized regression to remove systematic bias introduced by deterministic features such as GC-content and quantile normalization to correct for global distortions.

Subfields: Molecular Biology

View Paper: https://www.semanticscholar.org/paper/0e4ef9fe745155f4711d453fbb1e84a88d1042bf

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Molecular Biology

Citations in the corpus

Differential expression analysis of multifactor RNA-Seq experiments with respect to biological variation (2012, methodology): “For example, the use of offsets in the log-linear models can easily accommodate non-linear normalization procedures, including those based on quantile normalization GC-sequence content (61).”
A comprehensive evaluation of normalization methods for Illumina high-throughput RNA sequencing data analysis (2013, methodology): “are used against MiTF in order to lower its expression [33].”
Robustly detecting differential expression in RNA sequencing data using observation weights (2013, result): “As a positive control for this, we compared the observed weights to the sample-specific guanine-cytosine (GC) effects observed in the Pickrell dataset (10,35).”
limma powers differential expression analyses for RNA-sequencing and microarray studies (2015, methodology): “Alternatively, voom has the ability to respect normalization factors computed outside of limma by methods such as trimmed mean of M -values (41) or conditional quantile normalization (42).”
Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2 (2014, positive): “, using cqn [12] or EDASeq [13]), which may differ from gene to gene.”
Multi-stage differentiation defines melanoma subtypes with differential vulnerability to drug-induced iron-dependent oxidative stress (2018, methodology): “20) (Hansen et al., 2012) https://bioconductor.org/packages/release/bioc/html/cqn.html sigclust R Package (v1.”
GC-Content Normalization for RNA-Seq Data (2011, methodology): “[12], the GC-content bias on read counts is sample-specific, meaning that the dependence of gene counts on GC-content may vary between lanes.”
Count-based differential expression analysis of RNA sequencing data using R and Bioconductor (2013, methodology): “In general, we recommend starting from a sample metadata table that contains”
Normalization of RNA-seq data using factor analysis of control genes or samples (2014, background): “…last few years, several normalization strategies have been proposed to correct for between-sample distributional differences in read counts, such as differences in total counts (i.e., sequencing depths) 1,4 , and within-sample gene-specific effects, such as gene length or GC-content effects 2,5 .”
Cellular senescence mediates fibrotic pulmonary disease (2017, positive): “Conditional quantile normalization using the cqn Bioconductor package 48 was applied to the RNAseq data, to reduce variability introduced by GC content, gene size and total gene counts per sample.”