Simultaneous Testing of Grouped Hypotheses: Finding Needles in Multiple Haystacks

Cai, Tommaso; Sun, Wenguang

doi:10.1198/jasa.2009.tm08415

Cited by 107 publications

(113 citation statements)

References 23 publications

Supporting

Mentioning

112

Contrasting

Order By: Relevance

“…The advantage of this approach is that it uses the local FDR based on which optimal oracle procedures were developed for SMA (Sun and Cai 2007;Cai and Sun 2009).…”

Section: Mcp and Scadmentioning

confidence: 99%

“…Cai and Sun estimated the local FDR within each group and then applied the (approximate oracle) thresholding procedure to the combined set of local FDR values. For SMA of multiple chromosomes, we compared (i) pooled analysis based on the BH method or the local FDR thresholding procedure, (ii) separate analysis based on BH or the local FDR procedure, and (iii) the local FDR-based grouping procedure of Cai and Sun (2009). Sun and Cai (2009) extended their earlier method to dependent test statistics, using a hidden Markov model (HMM).…”

Section: Analyses Of Multiple Chromosomesmentioning

confidence: 99%

“…Sun and Cai (2009) extended their earlier method to dependent test statistics, using a hidden Markov model (HMM). Wei et al (2009) combined this method with the group-based method of Cai and Sun (2009) in the context of GWAS and implemented it in the R package PLIS. We tested Figure 1 Single-marker analysis (SMA) of chromosome 21 data with two isolated QTL, 200 replicates, and sample size N = 201.…”

Section: Analyses Of Multiple Chromosomesmentioning

confidence: 99%

“…0.05 for T = 0.25 and 0.3 but most were insignificantly higher (largest P-value of 0.001 for separate BH). Differences in TPR between SMA-based pooled and separate analyses and the grouping procedure of Cai and Sun (2009) were small and insignificant, although there was a trend for increasing TPR from pooled analysis to separate analysis to the grouping procedure (see Table S12 in File S1).…”

Section: Single-chromosome Simulation With Two/eight Qtl: Previous Apmentioning

confidence: 99%

See 3 more Smart Citations

Penalized Multimarkervs.Single-Marker Regression Methods for Genome-Wide Association Studies of Quantitative Traits

Breheny

Imam

et al. 2014

Genetics

View full text Add to dashboard Cite

The data from genome-wide association studies (GWAS) in humans are still predominantly analyzed using single-marker association methods. As an alternative to single-marker analysis (SMA), all or subsets of markers can be tested simultaneously. This approach requires a form of penalized regression (PR) as the number of SNPs is much larger than the sample size. Here we review PR methods in the context of GWAS, extend them to perform penalty parameter and SNP selection by false discovery rate (FDR) control, and assess their performance in comparison with SMA. PR methods were compared with SMA, using realistically simulated GWAS data with a continuous phenotype and real data. Based on these comparisons our analytic FDR criterion may currently be the best approach to SNP selection using PR for GWAS. We found that PR with FDR control provides substantially more power than SMA with genomewide type-I error control but somewhat less power than SMA with Benjamini-Hochberg FDR control (SMA-BH). PR with FDR-based penalty parameter selection controlled the FDR somewhat conservatively while SMA-BH may not achieve FDR control in all situations. Differences among PR methods seem quite small when the focus is on SNP selection with FDR control. Incorporating linkage disequilibrium into the penalization by adapting penalties developed for covariates measured on graphs can improve power but also generate more false positives or wider regions for follow-up. We recommend the elastic net with a mixing weight for the Lasso penalty near 0.5 as the best method.T HE goal of genome-wide association studies (GWAS) in humans and model organisms is to select a small subset of DNA markers, typically single-nucleotide polymorphisms (SNPs), which are in strong linkage disequilibrium (LD) with functional polymorphisms affecting a biomedical/clinical trait of interest. The selected markers are then replicated in other GWAS, fine-mapped, and further validated. GWAS may be viewed as a large-scale variable selection problem, with several millions of common SNPs, measured directly or imputed, available in current studies in humans.GWAS practitioners still strongly rely on single-marker analysis (SMA), including linear regression for continuous phenotypes, with control of the genome-wise error rate (GWER) [a special case of the family-wise error rate (FWER)], which accounts for the multiplicity of the entire genome. Assuming that a biomedical trait of interest is affected by multiple polymorphisms with "detectable" effects, SMA fits an incorrect model, and it is sensible to consider alternative methods that test all or subsets of markers simultaneously. This approach requires a form of penalized regression (PR) as the number of SNPs is much larger than the sample size.A major practical issue with the use of penalized regression is the determination of "optimal" values for the tuning parameter(s) and the lack of an error rate associated with the selection of SNPs. Common approaches for tuning parameter value determination include cross-validati...

show abstract

“…The advantage of this approach is that it uses the local FDR based on which optimal oracle procedures were developed for SMA (Sun and Cai 2007;Cai and Sun 2009).…”

Section: Mcp and Scadmentioning

confidence: 99%

Section: Analyses Of Multiple Chromosomesmentioning

confidence: 99%

Section: Analyses Of Multiple Chromosomesmentioning

confidence: 99%

Section: Single-chromosome Simulation With Two/eight Qtl: Previous Apmentioning

confidence: 99%

See 2 more Smart Citations

Penalized Multimarkervs.Single-Marker Regression Methods for Genome-Wide Association Studies of Quantitative Traits

Breheny

Imam

et al. 2014

Genetics

View full text Add to dashboard Cite

show abstract

“…Developing FWER control procedures specially for grouped hypotheses is sometimes necessary. 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%