Genome-Wide Significance Levels and Weighted Hypothesis Testing

Roeder, Kathryn; Wasserman, Larry

doi:10.1214/09-sts289

Cited by 108 publications

(159 citation statements)

References 39 publications

Supporting

Mentioning

156

Contrasting

Order By: Relevance

“…Filtering is analogous to the use of a common weight (m∕jMj) for all hypotheses passing the filter, and weight zero for the remainder. The use of continuously varying weights, on the other hand, has been shown to be optimal for certain experiment-wide definitions of type I error rate and power, and schemes for data-based estimation of these weights have been proposed (28,29). Our aim in this article, however, has not been to identify an optimal procedure, but rather to better understand filtering and to explore its effect on power and error rate control.…”

Section: Figmentioning

confidence: 99%

Independent filtering increases detection power for high-throughput experiments

Bourgon

Gentleman²,

Huber

2010

Proc. Natl. Acad. Sci. U.S.A.

643

633

View full text Add to dashboard Cite

With high-dimensional data, variable-by-variable statistical testing is often used to select variables whose behavior differs across conditions. Such an approach requires adjustment for multiple testing, which can result in low statistical power. A two-stage approach that first filters variables by a criterion independent of the test statistic, and then only tests variables which pass the filter, can provide higher power. We show that use of some filter/test statistics pairs presented in the literature may, however, lead to loss of type I error control. We describe other pairs which avoid this problem. In an application to microarray data, we found that gene-by-gene filtering by overall variance followed by a t-test increased the number of discoveries by 50%. We also show that this particular statistic pair induces a lower bound on fold-change among the set of discoveries. Independent filtering-using filter/ test pairs that are independent under the null hypothesis but correlated under the alternative-is a general approach that can substantially increase the efficiency of experiments.gene expression | multiple testing

show abstract

Section: Figmentioning

confidence: 99%

Independent filtering increases detection power for high-throughput experiments

Bourgon

Gentleman²,

Huber

2010

Proc. Natl. Acad. Sci. U.S.A.

643

633

View full text Add to dashboard Cite

show abstract

“…The commonly used criterion for error control in multiple testing is the family-wise error rate (FWER), which is the probability of rejecting at least one true null hypothesis (e.g., [20] ). The power of a multiple testing procedure can be measured by average power, taken to be the average power of individual tests corresponding to all false null hypotheses [21] .…”

Section: Multiple Testing For Genome-wide Association Studiesmentioning

confidence: 99%

A Generalized Sequential Bonferroni Procedure Using Smoothed Weights for Genome-Wide Association Studies Incorporating Information on Hardy-Weinberg Disequilibrium among Cases

et al. 2011

View full text Add to dashboard Cite

Background/Objectives: For genome-wide association studies (GWAS) with case-control designs, one of the most widely used association tests is the Cochran-Armitage (CA) trend test assuming an additive mode of inheritance. The CA trend test often has higher power than other association tests under additive and multiplicative disease models. However, it can have very low power under a recessive disease model in GWAS. Although tests (such as MAX3) robust to different genetic models have been developed, they often have relatively lower power than the CA trend test under additive and multiplicative models. The goal of this study is to propose an efficient method that not only has higher power than the CA trend test under dominant and recessive models but also maintains the power of the CA trend test under additive and multiplicative models. Methods: We employed the generalized sequential Bonferroni (GSB) procedure of Holm to incorporate information from a Hardy-Weinberg disequilibrium (HWD) test into the CA trend test based on estimating weights from the p values of the HWD test. We proposed to smooth the weights to reduce possible noise. Results and Conclusions: Results from extensive simulation studies showed that the proposed GSB procedure can achieve the goal described above.

show abstract

“…Efron (2004) and Cai and Sun (2009) pointed out that ignoring group labels may cause misleading results for separate groups, ''because highly significant cases from one group may be hidden among the nulls from another group'' (Cai and Sun, 2009). For grouped hypotheses, Hu et al (2010) and Cai and Sun (2009) proposed different procedures, but their goal is to control the false discovery rate (FDR); Roeder and Wasserman's (2009) procedure is aimed for the FWER control, however, the FWER control of their procedure was not shown when the weights are learned from the data.…”

Section: Introductionmentioning

confidence: 99%

“…In recent years there is an increasing interest in constructing testing procedures based on weighted p-values. For example, Benjamini and Hochberg (1997), Genovese et al (2006), Roeder and Wasserman (2009), Roquain and Van De Wiel (2009), Guo (2009) and Hu et al (2010. Gui et al (2012) reviewed the existing weighted p-value methods extensively.…”

Section: Introductionmentioning

confidence: 99%

“…Gui et al (2012) reviewed the existing weighted p-value methods extensively. The results in Genovese et al (2006) and Roeder and Wasserman (2009) showed that weighted p-value procedures have the attractive property of robustness to misspecification of the weights, i.e., well specified weights lead to huge power gains, but poorly specified weights lead to minute power loss. We also consider a weighed p-value procedure for grouped hypotheses to control the FWER.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation