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

Hypothesis weighting improves the power of large-scale multiple testing. We describe a method that uses covariates independent of the p-values under the null hypothesis, but informative of each test’s power or prior probability of the null hypothesis. Independent hypothesis weighting (IHW) increases power while controlling the false discovery rate (FDR). IHW is a practical approach to discover associations in large datasets as encountered in genomics and high-throughput biology. Availability: www.bioconductor.org/packages/IHW