Data-driven hypothesis weighting increases detection power in genome-scale multiple testing

By Nikolaos Ignatiadis, Bernd Klaus, Judith B. Zaugg, Wolfgang Huber. Nature Methods, 2016

595 citations recorded by the source corpus.

Hypothesis weighting improves the power of large-scale multiple testing. We describe independent hypothesis weighting (IHW), a method that assigns weights using covariates independent of the P-values under the null hypothesis but informative of each test's power or prior probability of the null hypothesis (http://www.bioconductor.org/packages/IHW). IHW increases power while controlling the false discovery rate and is a practical approach to discovering associations in genomics, high-throughput biology and other large data sets.

Subfields: Molecular Biology, Statistics and Probability

View Paper: https://www.semanticscholar.org/paper/1ee765d218aaea7ef4d95df27043954fb7d3b4df

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

In Statistics and Probability

Citations in the corpus

Splitting strategies for post-selection inference (2021, methodology): “Methods based on data splitting have also been considered in more recent works such as Rubin et al. (2006), Wasserman and Roeder (2009), Ignatiadis et al. (2016), and DiCiccio et al. (2020).”
CHiCAGO: robust detection of DNA looping interactions in Capture Hi-C data (2015, methodology): “These approaches have been used successfully to incorporate prior knowledge in GWAS [33–35] and are emerging in functional genomics analyses [36, 37].”
Leveraging polygenic functional enrichment to improve GWAS power (2017, methodology): “We compared FINDOR to three other methods that can incorporate auxiliary information for each SNP: stratified false discovery rate (S-FDR),(27) grouped Benjamini Hochberg (GBH),(28) and independent hypothesis weighting (IHW).(29) For completeness, we also compared to GenoWAP,(23) a Bayesian integrat”
RNA Sequencing Data: Hitchhiker's Guide to Expression Analysis (2018, background): “Storey's qvalue, for instance, estimates the proportion of true null hypotheses from the data to increase power [158] , while others adopt a datadriven weighting of the pvalues in the FDR correction [159] .”
Universal Patterns of Selection in Cancer and Somatic Tissues (2018, unknown): “Alternative approaches that can help increase power on a priori gene candidates without theoretically incurring in an inflation of the global false discovery rate are stratified false discovery rate (Sun et al., 2006) and data-driven hypothesis weighting (Ignatiadis et al., 2016).”
Covariate powered cross‐weighted multiple testing (2017, methodology): “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).”