svaseq: removing batch effects and other unwanted noise from sequencing data

By Jeffrey T. Leek. bioRxiv, 2014

513 citations recorded by the source corpus.

It is now well known that unwanted noise and unmodeled artifacts such as batch effects can dramatically reduce the accuracy of statistical inference in genomic experiments. We introduced surrogate variable analysis for estimating these artifacts by (1) identifying the part of the genomic data only affected by artifacts and (2) estimating the artifacts with principal components or singular vectors of the subset of the data matrix. The resulting estimates of artifacts can be used in subsequent analyses as adjustment factors. Here I describe an update to the sva approach that can be applied to analyze count data or FPKMs from sequencing experiments. I also describe the addition of supervised sva (ssva) for using control probes to identify the part of the genomic data only affected by artifacts. These updates are available through the surrogate variable analysis (sva) Bioconductor package.

Subfields: Cancer Research, Genetics

View Paper: https://www.semanticscholar.org/paper/0411079e9c554ae07970b640b4a0371f21c7beec

How researchers see this paper

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

In Cancer Research

In Genetics

Citations in the corpus

Developmental Relationships of Four Exhausted CD8+ T Cell Subsets Reveals Underlying Transcriptional and Epigenetic Landscape Control Mechanisms (2020, methodology): “(Leek, 2014) Correlation network analysis—”
MAST: a flexible statistical framework for assessing transcriptional changes and characterizing heterogeneity in single-cell RNA sequencing data (2015, methodology): “Finally, we investigated the relationship between our approach and the weight correction of Shalek et al. [5] and other technical bias correction approaches like Re-move Unwanted Variation (RUV) and Surrogate Variable Analysis (SVA; Additional file 1: Figure S5A, B) [17, 18].”
RNA Sequencing Data: Hitchhiker's Guide to Expression Analysis (2018, background): “SVA [155,156] and RUV [125], for instance, estimate surrogate variables through singular value decomposition on control features or on a matrix of model residuals so as to avoid that the phenotypic effect of interest is captured by the surrogates.”
Human Tumor-Associated Macrophage and Monocyte Transcriptional Landscapes Reveal Cancer-Specific Reprogramming, Biomarkers, and Therapeutic Targets (2019, methodology): “Normalized readswere fed to the Combat function of the Surrogate Variable Analysis (SVA) package (Leek, 2014).”
ComBat-seq: batch effect adjustment for RNA-seq count data (2020, methodology): “For heterogeneity from unknown sources, SVASeq (5) and RUVSeq (3) are commonly used.”
Single-Cell RNA-Seq Technologies and Related Computational Data Analysis (2019, background): “Because of the data feature diﬀerences between scRNA-seq and bulk RNA-seq, batch-correction approaches specially proposed for bulk RNA-seq [e.g., RUVseq (Risso et al., 2014) and svaseq (Leek, 2014)] may not be suitable for scRNA-seq.”
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.”
Integrative single-cell analysis (2019, background): “This challenge echoes similar needs for ‘batch-correction’ techniques for bulk data sets, which are essential for data sets produced across different laboratories and experimental workflows 79 .”
Normalizing single-cell RNA sequencing data: Challenges and opportunities (2017, background): “Additionally, in any of these high-throughput assays, one needs to account for possibly more complex and putatively unknown effects, collectively known as “batch effects” [18–20].”