Estimating the proportion of true null hypotheses when the statistics are discrete

Dialsingh, Isaac; Austin, Stefanie; Altman, Naomi

doi:10.1093/bioinformatics/btv104

Cited by 16 publications

(18 citation statements)

References 21 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In addition, we will not consider estimators in Dialsingh et al. () since they are based on the two groups model for the p values.…”

Section: Simulation Studymentioning

confidence: 99%

“…However, we will not investigate the estimator̂P C * 0 = min{1, −1 ∑ =1 −1 } proposed by Pounds and Cheng (2006), where is the mean of computed under the null hypothesis, since we have observed in Chen and Doerge (2014) that̂P C * 0 is usually 1 when 0 ≥ 0.5 for a similar simulation setup (see Section 4.1 for the simulation design). In addition, we will not consider estimators in Dialsingh et al (2015) since they are based on the two groups model for the p values.…”

Section: Simulation Studymentioning

confidence: 99%

“…There has been evidence that the BH procedure and Storey's procedure tend to be less powerful or may yield unreliable results when applied to the discrete paradigm; see, for example, Gilbert () and Pounds and Cheng (). To develop better FDR procedures for the discrete paradigm, three major approaches have been taken: (i) modify the step‐up sequence in the BH procedure according to the achievable significance level of a discrete p ‐value distribution; see, for example, Tarone (), Gilbert (), and Heyse (); (ii) use randomized p values or mid p values; see, for example, Kulinskaya and Lewin (), Heller and Gur (), and Habiger (); and (iii) propose less conservative estimators of the proportion π 0 of true null hypotheses and use them to induce more powerful adaptive FDR procedures, see, for example, Benjamini, Krieger, and Yekutieli (), Pounds and Cheng (), Blanchard and Roquain (), Chen and Doerge (), Liang (), and Dialsingh, Austin, and Altman ().…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Multiple testing with discrete data: Proportion of true null hypotheses and two adaptive FDR procedures

2018

View full text Add to dashboard Cite

We consider multiple testing with false discovery rate (FDR) control when p values have discrete and heterogeneous null distributions. We propose a new estimator of the proportion of true null hypotheses and demonstrate that it is less upwardly biased than Storey's estimator and two other estimators. The new estimator induces two adaptive procedures, that is, an adaptive Benjamini-Hochberg (BH) procedure and an adaptive Benjamini-Hochberg-Heyse (BHH) procedure. We prove that the adaptive BH (aBH) procedure is conservative nonasymptotically. Through simulation studies, we show that these procedures are usually more powerful than their nonadaptive counterparts and that the adaptive BHH procedure is usually more powerful than the aBH procedure and a procedure based on randomized p-value. The adaptive procedures are applied to a study of HIV vaccine efficacy, where they identify more differentially polymorphic positions than the BH procedure at the same FDR level.

show abstract

“…In addition, we will not consider estimators in Dialsingh et al. () since they are based on the two groups model for the p values.…”

Section: Simulation Studymentioning

confidence: 99%

Section: Simulation Studymentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Multiple testing with discrete data: Proportion of true null hypotheses and two adaptive FDR procedures

2018

View full text Add to dashboard Cite

show abstract

“…By filtering out low power tests ( i.e. low count regions) with the T method [32], the p-value distribution becomes more uniform and the p-values can be adjusted for multiple testing. Filtered p-values are then transformed to q-values [33].…”

Section: Resultsmentioning

confidence: 99%

normR: Regime enrichment calling for ChIP-seq data

Helmuth

Arrigoni

et al. 2016

Preprint

View full text Add to dashboard Cite

ChIP-seq probes genome-wide localization of DNA-associated proteins. To mitigate technical biases ChIP-seq read densities are normalized to read densities obtained by a control. Our statistical framework “normR” achieves a sensitive normalization by accounting for the effect of putative protein-bound regions on the overall read statistics. Here, we demonstrate normR’s suitability in three studies: (i) calling enrichment for high (H3K4me3) and low (H3K36me3) signal-to-ratio data; (ii) identifying two previously undescribed H3K27me3 and H3K9me3 heterochromatic regimes of broad and peak enrichment; and (iii) calling differential H3K4me3 or H3K27me3-enrichment between HepG2 hepatocarcinoma cells and primary human Hepatocytes. normR is readily available on http://bioconductor.org/packages/normr

show abstract

“…The FDR is less restrictive and allows a certain amount of false positives [6]. A number of methods have been established for estimation of the proportion of null hypotheses when the test statistics are discrete [7]. In this paper, we compare 12 methods for comparing false discovery rates when the test statistics are continuous under varying dependence structures.…”

Section: Notationmentioning

confidence: 99%

Comparison of Methods for Estimating the Proportion of Null Hypotheses π0 in High Dimensional Data When the Test Statistics is Continuous

Dialsingh¹,

Cedeno²

2017

J Biom Biostat

Self Cite

View full text Add to dashboard Cite

Advances in Genomics have re-energized interest in multiple hypothesis testing procedures but have simultaneously created new methodological and computational challenges. In Genomics for instance, it is now commonplace for experiments to measure expression levels in thousands of genes creating large multiplicity problems when thousands of hypotheses are to be tested simultaneously. Within this context we seek to identify differentially expressed genes, that is, genes whose expression levels are associated with a particular response or covariate of interest. The False Discovery Rate (FDR) is the preferred measure since the Family Wise Error Rates (FWERs) are usually overly restrictive. In the FDR methods, estimation of the proportion of null hypotheses (π0) is an important parameter that needs to be estimated.In this paper, we compare the effectiveness of 12 methods for estimating π0 when the test statistics are continuous using simulated data with independent, weak dependence, and moderate dependence structures.

show abstract

Estimating the proportion of true null hypotheses when the statistics are discrete

Abstract: implemented in R.

Cited by 16 publications

References 21 publications

Multiple testing with discrete data: Proportion of true null hypotheses and two adaptive FDR procedures

Multiple testing with discrete data: Proportion of true null hypotheses and two adaptive FDR procedures

normR: Regime enrichment calling for ChIP-seq data

Comparison of Methods for Estimating the Proportion of Null Hypotheses π0 in High Dimensional Data When the Test Statistics is Continuous

Contact Info

Product

Resources

About