## False Discovery Rates and Copy Number Variation

### Cached

### Download Links

Citations: | 3 - 2 self |

### BibTeX

@MISC{Efron_falsediscovery,

author = {Bradley Efron and Nancy Zhang},

title = {False Discovery Rates and Copy Number Variation},

year = {}

}

### OpenURL

### Abstract

Copy number changes, the gains and losses of chromosome segments, are a common type of genetic variation among healthy individuals as well as an important feature in tumor genomes. Microarray technology enables us to simultaneously measure, with moderate accuracy, copy number variation at more than a million chromosome locations and for hundreds of subjects. This leads to massive data sets and complicated inference problems concerning which locations for which subjects are genuinely variable. In this paper we consider a relatively simple false discovery rate approach to cnv analysis. More careful parametric change-point methods can then be focused on promising regions of the genome. Key words and phrases: copy number 1

### Citations

2771 |
Estimating the dimension of a model
- Schwarz
- 1978
(Show Context)
Citation Context ...1, . . . , τm (m is unknown). The data is assumed Gaussian, with the mean shifting at each change-point, but with the variance remaining constant. While Yao (1988) showed that the traditional BIC 24(=-=Schwarz, 1978-=-) is consistent for the estimation of m, Zhang and Siegmund (2007) showed that it is not consistent in estimating the Bayes factor, the quantity that underlies the classic BIC. Zhang and Siegmund (200... |

286 | The jackknife and the bootstrap for general stationary observations
- Künsch
- 1989
(Show Context)
Citation Context ...lot shows the ˆ ki estimates, plotted in the positive direction for gains and plotted inverted in the negative direction for losses. Here we employed the block bootstrap method on the {tdrij} matrix (=-=Künsch, 1989-=-) to find the distribution of k max = max ˆki i (a different calculation from, but in similar spirit to, the one used in Figure 5). Briefly, N/L blocks of size L are sampled with replacement from subj... |

265 | High resolution analysis of dna copy number variation using comparative genomic hybridization to microarrays - Pinkel, Segraves, et al. - 1998 |

136 |
Circular binary segmentation for the analysis of array-based DNA copy number data
- OLSHEN, VENKATRAMAN, et al.
- 2004
(Show Context)
Citation Context ...mple analysis and the simultaneous processing of multiple samples, global change-point tests, scanning over the entire range of positions, have played a central role in the statistical cnv lierature (=-=Olshen et al., 2004-=-; Siegmund, Yakir and Zhang, 2010; Zhang et al., 2010). The literature leans heavily on Gaussian process theory, and within that realm produces impressively precise testing algorithms. Wang et al. (20... |

105 | Genome-wide analysis of DNA copy-number changes using cDNA microarrays - POLLACK, PEROU, et al. - 1999 |

86 | An empirical Bayes approach to statistics - Robbins - 1956 |

85 | A comparison study: applying segmentation to arrayCGH data for downstream analyses - WILLENBROCK, FRIDLYAND - 2005 |

81 | Assembly of microarrays for genome-wide measurement of dna copy - Snijders, Nowak, et al. - 2001 |

71 | Comparative analysis of algorithms for identifying amplifications and deletions in array CGH data - LAI, JOHNSON, et al. - 2005 |

69 | Defining the curvature of a statistical problem (with applications to second order efficiency), Ann - Efron - 1975 |

50 | A high-resolution survey of deletion polymorphism in the human genome - Conrad, Andrews, et al. - 2006 |

44 | Spatial smoothing and hot spot detection for CGH data using the fused lasso - Tibshirani, Wang - 2008 |

40 | Estimating the number of change-points via Schwarz’ criterion - Yao - 1988 |

36 |
Assessing the significance of chromosomal aberrations in cancer: methodology and applicaCancer Cell NF1 Inactivation in Gliomagenesistion to glioma
- Beroukhim, Getz, et al.
- 2007
(Show Context)
Citation Context ...dr approach, closer to the methods proposed here. The identification of cnv-prone regions across a cohort of tumor samples has been a problem of increased scientific interest. Most published methods (=-=Beroukhim et al., 2007-=-; Diskin et al., 2006; Guttman et al., 2007; Newton et al., 1998; Newton and Lee, 2000; Rouveirol et al., 2006; Taylor et al., 2008) take a post-segmentation approach: each sample is first segmented i... |

30 |
STAC: A method for testing the significance of DNA copy number aberrations across multiple array-CGH experiments
- Diskin
- 2006
(Show Context)
Citation Context ...he methods proposed here. The identification of cnv-prone regions across a cohort of tumor samples has been a problem of increased scientific interest. Most published methods (Beroukhim et al., 2007; =-=Diskin et al., 2006-=-; Guttman et al., 2007; Newton et al., 1998; Newton and Lee, 2000; Rouveirol et al., 2006; Taylor et al., 2008) take a post-segmentation approach: each sample is first segmented individually, which re... |

30 | Microarrays, empirical Bayes and the two-groups model
- Efron
- 2008
(Show Context)
Citation Context ...values renders copy number variations far more visible; see Figure 3. The paper develops as follows: an iterative algorithm is introduced in Section 2, in which a local false discovery rate estimate (=-=Efron, 2008-=-) is first fit to the combined data, and then modified to take account of differing cnv probabilities at the various positions i. This gives an fdr estimate for each position and subject, as well as a... |

25 | Large-Scale Inference: Empirical Bayes Methods for Estimation, Testing and Prediction - Efron - 2010 |

22 | High- resolution analysis of DNA copy number using oligonucleotide microarrays - BIGNELL, HUANG, et al. - 2004 |

19 |
Computation of recurrent minimal genomic alterations from array-CGH data
- Rouveirol
- 2006
(Show Context)
Citation Context ...umor samples has been a problem of increased scientific interest. Most published methods (Beroukhim et al., 2007; Diskin et al., 2006; Guttman et al., 2007; Newton et al., 1998; Newton and Lee, 2000; =-=Rouveirol et al., 2006-=-; Taylor et al., 2008) take a post-segmentation approach: each sample is first segmented individually, which reduces them to piece-wise constant sequences indicating regions of amplification, deletion... |

16 | Highresolution genomic profiling of chromosomal aberrations using Infinium whole-genome genotyping - Peiffer, Le, et al. - 2006 |

13 |
A modified Bayes information criterion with applications to comparative genomic hybridization data
- ZHANG, SIEGMUND
- 2006
(Show Context)
Citation Context ...values (the xij’s). The red points are locations with fdr < 0.05, and the black points are locations with fdr < 0.005. The vertical lines show the change-points determined by the modified BIC method (=-=Zhang and Siegmund, 2007-=-). Under the above model, the generalized likelihood ratio assuming known τ1, τ2 and maximized over µ has the form τ2−1 ∑ / L(τ1, τ2) = yi (τ2 − τ1). i=τ1 Maximizing over τ1 and τ2, the generalized li... |

11 | On the statistical analysis of allelic-loss data
- Newton, Gould, et al.
- 1998
(Show Context)
Citation Context ...n of cnv-prone regions across a cohort of tumor samples has been a problem of increased scientific interest. Most published methods (Beroukhim et al., 2007; Diskin et al., 2006; Guttman et al., 2007; =-=Newton et al., 1998-=-; Newton and Lee, 2000; Rouveirol et al., 2006; Taylor et al., 2008) take a post-segmentation approach: each sample is first segmented individually, which reduces them to piece-wise constant sequences... |

10 | Sequence and structural variation in a human genome uncovered by short-read, massively parallel ligation sequencing using two-base encoding. Genome research 19 - McKernan - 2009 |

10 | Simultaneous inference: When should hypothesis testing problems be combined
- EFRON, B
- 2008
(Show Context)
Citation Context ...Null Mixture Non-null • Local false discovery rate fdr(z) = Pr{null|z} = πo f0(z)/ f (z) • Empirical Bayes z −→ ˆπ0, ˆ f0, ˆ f −→ ̂ fdr(z) = ˆπ0 ˆ f0(z) / ˆ f (z) • Reject H0m if ̂ fdr(zm) small (see =-=Efron, 2008-=-) FDRs and CNV 7Estimated local false discovery rate, all 750,000 z−values; pihat0=.954, estimated null density N(.04,.93^2) local fdr 0.0 0.2 0.4 0.6 0.8 1.0 −3.30 3.57 −6 −4 −2 0 2 4 6 z value fdrh... |

8 |
Assessing the significance of conserved genomic aberrations using high resolution genomic microarrays. PLoS Genet 3: e143
- Guttman, Mies, et al.
- 2007
(Show Context)
Citation Context ...ere. The identification of cnv-prone regions across a cohort of tumor samples has been a problem of increased scientific interest. Most published methods (Beroukhim et al., 2007; Diskin et al., 2006; =-=Guttman et al., 2007-=-; Newton et al., 1998; Newton and Lee, 2000; Rouveirol et al., 2006; Taylor et al., 2008) take a post-segmentation approach: each sample is first segmented individually, which reduces them to piece-wi... |

7 | Correlated z-values and the accuracy of large-scale statistical estimates - Efron - 2010 |

7 |
Tail probabilities for the maxima of some random fields
- Siegmund
- 1988
(Show Context)
Citation Context ...e xij’s). The red points are locations with fdr < 0.05, and the black points are locations with fdr < 0.005. The vertical lines show the change-points determined by the modified BIC method (Zhang and =-=Siegmund, 2007-=-). Under the above model, the generalized likelihood ratio assuming known τ1, τ2 and maximized over µ has the form τ2−1 ∑ / L(τ1, τ2) = yi (τ2 − τ1). i=τ1 Maximizing over τ1 and τ2, the generalized li... |

6 |
Inferring the location and effect of tumor suppressor genes by instability-selection modeling of allelic-loss data
- Newton, Lee
- 2000
(Show Context)
Citation Context ...s across a cohort of tumor samples has been a problem of increased scientific interest. Most published methods (Beroukhim et al., 2007; Diskin et al., 2006; Guttman et al., 2007; Newton et al., 1998; =-=Newton and Lee, 2000-=-; Rouveirol et al., 2006; Taylor et al., 2008) take a post-segmentation approach: each sample is first segmented individually, which reduces them to piece-wise constant sequences indicating regions of... |

4 | Optimal and fast detection of spatial clusters with scan statistics - Walther - 2010 |

3 |
Functional copy-number alterations in cancer. PLoS One 3: e3179
- BS, Barretina, et al.
- 2008
(Show Context)
Citation Context ...problem of increased scientific interest. Most published methods (Beroukhim et al., 2007; Diskin et al., 2006; Guttman et al., 2007; Newton et al., 1998; Newton and Lee, 2000; Rouveirol et al., 2006; =-=Taylor et al., 2008-=-) take a post-segmentation approach: each sample is first segmented individually, which reduces them to piece-wise constant sequences indicating regions of amplification, deletion, or normal copy numb... |

3 | A method for calling gains and losses in array-CGH data - WANG, KIM, et al. - 2005 |

2 | A bayes regression approach to array-cgh data. Statistical Applications in Molecular Biology 5 - Wen, Wu, et al. - 2006 |

1 | Empirical bayes estimates for large-scale prediction problems - unknown authors - 2009 |

1 | An initial map of insertion and deletion (indel) variation in the human genome. Genome Res 16 - Devine - 2006 |

1 | Detecting simultaneous variant intervals in aligned sequences, submitted - Siegmund, Yakir, et al. - 2010 |

1 | The cancer genome. Nature 458: 719–724 - Stratton, Campbell, et al. - 2009 |

1 | DNA Copy Number Profiling in Normal and Tumor Genomes - unknown authors - 2010 |

1 | FDRs and CNV 25 Tibshirani - Wang, Kim, et al. - 2005 |