Independent filtering increases detection power for high-throughput experiments

By Richard Bourgon, Robert Gentleman, Wolfgang Huber. Proceedings of the National Academy of Sciences of the United States of America, 2010

764 citations recorded by the source corpus.

Subfields: Molecular Biology

View Paper: https://www.semanticscholar.org/paper/8c1eed72bb4c630e18c051b062dfab77e12eddb6

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

MetaboAnalyst 2.0—a comprehensive server for metabolomic data analysis (2012, background): “This filtering step can usually improve the performance and reduce the false discovery rates (FDR) in downstream statistical analysis (12,13).”
Cell-to-cell expression variability followed by signal reinforcement progressively segregates early mouse lineages (2013, methodology): “tex", open="w") > for(i in seq(along = costs)) { + sxqdsub = switch(i, + { + rv = sxqd[, sxqd$sampleGroup %in% groups[2] ] + exprs(rv)[ exprs(rv)>2 ] = 2 + rv + }, { + rv = sxqd[, sxqd$sampleGroup %in% groups[3] ] + exprs(rv)[ exprs(rv)<2 ] = 2 + exprs(rv) = exprs(rv)-1”
RNA Sequencing Data: Hitchhiker's Guide to Expression Analysis (2018, background): “[115] showed that filtering is valid as long as it is independent of the DE test statistic; thus, filtering on residual variance is invalid, while filtering on general expression strength, as is commonly done, is valid.”
Covariate powered cross‐weighted multiple testing (2017, neutral): “Using it as a covariate can lead to a large power increase in simultaneous two-sample t-tests (Bourgon et al., 2010; Ignatiadis et al., 2016).”
RNA-Seq workflow: gene-level exploratory analysis and differential expression (2015, neutral): “The independent filtering software used inside DESeq2 comes from the genefilter package, that contains a reference to a paper describing the statistical foundation for independent filtering 22 .”
Expression Atlas update—an integrated database of gene and protein expression in humans, animals and plants (2015, methodology): “The differential analysis is now performed using DESeq2 (18) with independent filtering (19).”
GSVA: gene set variation analysis for microarray and RNA-Seq data (2013, background): “Non-specific filtering of genes in high-throughput experiments has been shown to increase the statistical power to detect significant changes in gene expression levels [64] and this observation is likely to hold at gene set level.”
Comparison of software packages for detecting differential expression in RNA-seq studies (2013, methodology): “This can be done using, for instance, the overall sum of counts independent of the sample groups [16, 25] or the maximum normalized count across samples [26].”
Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2 (2014, neutral): “However, the loss can be reduced if genes are omitted from the testing that have little or no chance of being detected as differentially expressed, provided that the criterion for omission is independent of the test statistic under the null [21] (see Methods).”
Count-based differential expression analysis of RNA sequencing data using R and Bioconductor (2013, neutral): “In either case, the analysis for the current data set highlights that library type (single-end or paired-end) has a systematic effect on the read counts and provides an example of a data-driven modeling decision: here, a GLM-based analysis that accounts for the (assumed linear) effect of library…”
Data-driven hypothesis weighting increases detection power in genome-scale multiple testing (2016, background): “The covariate can be any continuous-valued or categorical variable that is thought to be informative on the statistical properties of the hypothesis tests, while it is independent of the p-value under the null hypothesis [9].”
Orchestrating single-cell analysis with Bioconductor (2019, background): “Furthermore, genes can be annotated with their abundance measurements, such as average expression, or frequency of detection, enabling the removal of uninformative genes [58].”