It is clear, however, that evidence of an excess of amino acid substitutions (at least from site - by - site tests) is no longer a sufficiently convincing demonstration of selection, not only because a high ratio could result from selection on synonymous mutations rather than positive selection on proteins [5], but also because there is potentially a high
false discovery rate of selected sites [6].
In order to control for type 1 error when conducting multiple comparisons, Benjamini and Hochberg's (1995)
rough false discovery rate was used and the level of significance was measured at below 0.025.
Here, we present a computational strategy for detecting «differentially abundant» populations by assigning cells to hyperspheres, testing for significant differences between conditions and controlling the
spatial false discovery rate.
Rows correspond to the putative miR - 142 and miR - 150 targets (
local false discovery rate (FDR) 1 %), and columns represent individual experimental samples.
Throughout, Q values indicate significance of differences after adjusting for multiple comparisons by controlling
the false discovery rate for selected comparisons.
These include methods «that emphasize estimation over testing such as confidence, credibility, or prediction intervals; Bayesian methods; alternative measures of evidence such as likelihood ratios or Bayes factors; and other approaches such as decision - theoretic modeling and
false discovery rates.»
Each of these metabolites was found to have
false discovery rates (FDRs) of less than 10 % (SI Appendix, Table S1 A and B).
Heat maps showing different gene expression patterns (color scale represents the fold change) using gene ontology (GO) analyses to compare the transcriptomes of the rope versus gill (A; P < 0.05,
false discovery rate ≤ 0.1) or muscle tissues (B; P < 0.05, false discovery rate ≤ 0.1) in mature male sea lamprey.
The cutoff for significance in these experiments was set at
a false discovery rate (FDR) of approximately 5 %.
We discuss common correction schemes such as Bonferroni, Holm, Benjamini & Hochberg and Storey's q and show how they impact the false positive rate (FPR),
false discovery rate (FDR) and power of a batch of tests.
A false discovery rate (FDR) correction was used to control for multiple comparisons.
These genes include synaptojanin 2 (SYNJ2), which has been detected with PAML [phylogenetic analysis by maximum likelihood;
false discovery rate (FDR) = 5 %] and represents a longevity gene candidate associated with variation in levels of agreeableness and cognitive abilities (60, 61).
We apply three different methods of multiple test correction, including Bonferroni,
false discovery rate, and permutations.
In that paper I was interested in the false positive rate (also known as
the false discovery rate) in tests of significance.
Among the 19,579 genes that are expressed in either tissue, 63.14 % (12,363) are differentially expressed with
a false discovery rate (FDR) of less than 5 %.
Genes were considered to be differentially expressed if they had an adjusted p - value of 0.05 or less (equivalent to
a false discovery rate of 5 %).
ALE meta - analysis: controlling
the false discovery rate and performing statistical contrasts.
In these analyses, we identify and subsequently analyze a set of 107 autosomal genes with
a false discovery rate (FDR) of < 30 %; in total, this larger set of genes harbor de novo loss of function (LoF) mutations in 5 % of cases, and numerous de novo missense and inherited LoF mutations in additional cases.
Controlling
the false discovery rate: A practical and powerful approach to multiple testing.
In contrast to Ex-miRNA, none of the miRNA in total plasma remained statistically significant if
the false discovery rate was below 15 percent.
A probe and the containing promoter were called differentially methylated if the p - value of the probe t - statistic was at most 0.05 (uncorrected for multiple testing), log2-fold change between the groups was at least 0.25, and
the false discovery rates (FDR) of the promoter - level statistic was at most 0.2.
Bar heights are - log10 values of
the false discovery rates, so bars higher than the dashed line have false discovery rates below 0.05.
Our use of
a false discovery rate of 0.2 as a threshold for calling differential methylation implies that we expect at most 20 % of our calls to be erroneous.
As a number of regressions were completed,
the false discovery rate method was used to control for potential type I error (Benjamini & Hockberg, 1995).