Normalization of RNA-seq data using factor analysis of control genes or samples

By Davide Risso, John Ngai, Terence P. Speed, Sandrine Dudoit. Nature Biotechnology, 2014

1,897 citations recorded by the source corpus.

Normalization of RNA-sequencing (RNA-seq) data has proven essential to ensure accurate inference of expression levels. Here, we show that usual normalization approaches mostly account for sequencing depth and fail to correct for library preparation and other more complex unwanted technical effects. We evaluate the performance of the External RNA Control Consortium (ERCC) spike-in controls and investigate the possibility of using them directly for normalization. We show that the spike-ins are not reliable enough to be used in standard global-scaling or regression-based normalization procedures. We propose a normalization strategy, called remove unwanted variation (RUV), that adjusts for nuisance technical effects by performing factor analysis on suitable sets of control genes (e.g., ERCC spike-ins) or samples (e.g., replicate libraries). Our approach leads to more accurate estimates of expression fold-changes and tests of differential expression compared to state-of-the-art normalization methods. In particular, RUV promises to be valuable for large collaborative projects involving multiple laboratories, technicians, and/or sequencing platforms.

Subfields: Molecular Biology

View Paper: https://www.semanticscholar.org/paper/73c42031d99e6f2d48d33969b281f84cffce5b47

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

Translating RNA sequencing into clinical diagnostics: opportunities and challenges (2016, background): “Furthermore, multiple groups have examined the role of normalization methods towards the controlling biases resulting from GC content, sequencing coverage and insert size 103, 104 .”
A deep generative model for single-cell RNA sequencing with application to detecting differentially expressed genes (2017, background): “Here the constant γn are optional covariates that can be passed to fw that account for batch effects, like normalization [12, 13], to remove unwanted variation from the latent representation.”
Selecting between‐sample RNA‐Seq normalization methods from the perspective of their assumptions (2016, methodology): “The two assumptions listed above indicate slightly diﬀerent things for the diﬀerent sub-methods, and RUVr doesn’t actually require controls (it is an adaptation of the RUV method to be used when controls are not available) [33].”
A deep generative model for gene expression profiles from single-cell RNA sequencing (2017, background): “Here the constant γn are optional covariates that can be passed to fw that account for batch effects, like normalization [12, 13], to remove unwanted variation from the latent representation.”
Orchestrating single-cell analysis with Bioconductor (2019, methodology): “Though many normalization procedures exist for data derived from traditional bulk assays [3, 7, 9, 60–62], using data derived from single-cell assays, such as scRNA-seq, introduces new challenges that require cell and gene-specific biases to be eliminated prior to downstream analyses that depend on ”
Controlling bias and inflation in epigenome- and transcriptome-wide association studies using the empirical null distribution (2016, methodology): “For TWAS, we additionally used RUV [27, 32] and SVA [26].”
MetaboAnalyst 5.0: narrowing the gap between raw spectra and functional insights (2021, methodology): “The utility function for batch effect and signal drift correction has been updated with eight algorithms-EigenMS (52), QC-RLSC (53), ANCOVA (54), RUV-random (55), RUV2 (56), RUVseq (57), NOMIS (58) and CCMS (59) for correction based on either the data itself, QC samples or internal standards.”
Quality control, modeling, and visualization of CRISPR screens with MAGeCK-VISPR (2015, methodology): “Existing batch removal algorithms, including ComBat [33] and RUVseq [34], have been widely used to remove batch effects in gene expression analysis.”
RNA-Seq workflow: gene-level exploratory analysis and differential expression (2015, positive): “Another package for detecting hidden batches is the RUVSeq package 25 , with the acronym “Remove Unwanted Variation”. library( " sva " ) Below we obtain a matrix of normalized counts for which the average count across samples is larger than 1.”
edgeR v4: powerful differential analysis of sequencing data with expanded functionality and improved support for small counts and larger datasets (2024, positive): “Observation-specific library sizes can be used to implement non-linear normalization in edgeR [ 11 , 12 , 27–29 ] or to allow for transcript length effects [ 30 ]. edgeR v4 also allows observation-specific dispersions by where the w gi are known GLM weights.”
A practical guide to single-cell RNA-sequencing for biomedical research and clinical applications (2017, background): “Several computational approaches have been devised that account for such variability [59, 70, 71].”
A Unified Model for the Function of YTHDF Proteins in Regulating m6A-Modified mRNA (2020, methodology): “To remove unwanted variation, the RUVg estimation package was used (Risso et al., 2014).”
Araport11: a complete reannotation of the Arabidopsis thaliana reference genome (2016, methodology): “We used RUVseq (Peixoto et al. 2015; Risso et al. 2014) to remove technical effects to accommodate the heterogenous sources of the RNA-seq datasets.”
A general and flexible method for signal extraction from single-cell RNA-seq data (2017, positive): “Other more complex, non-linear effects, such as those collectively known as batch effects, require more sophisticated normalization [28].”
Batch effects in single-cell RNA sequencing data are corrected by matching mutual nearest neighbours (2018, methodology): “Other methods such as RUVseq [9] and svaseq [10] are also frequently used for batch correction, but focus primarily on identifying unknown factors of variation, e.”