LowPass Copy number analysis (GISTIC2)
Brain Lower Grade Glioma (Primary solid tumor)
15 July 2014  |  analyses__2014_07_15
Maintainer Information
Citation Information
Maintained by Spring Yingchun Liu (Broad Institute)
Cite as Broad Institute TCGA Genome Data Analysis Center (2014): LowPass Copy number analysis (GISTIC2). Broad Institute of MIT and Harvard. doi:10.7908/C18S4NPT
Overview
Introduction

GISTIC identifies genomic regions that are significantly gained or lost across a set of tumors. The pipeline first filters out normal samples from the segmented copy-number data by inspecting the TCGA barcodes and then executes GISTIC version 2.0.21 (Firehose task version: 127).

Summary

There were 52 tumor samples used in this analysis: 12 significant arm-level results, 3 significant focal amplifications, and 2 significant focal deletions were found.

Results
Focal results

Figure 1.  Genomic positions of amplified regions: the X-axis represents the normalized amplification signals (top) and significance by Q value (bottom). The green line represents the significance cutoff at Q value=0.25.

Table 1.  Get Full Table Amplifications Table - 3 significant amplifications found. Click the link in the last column to view a comprehensive list of candidate genes. If no genes were identified within the peak, the nearest gene appears in brackets.

Cytoband Q value Residual Q value Wide Peak Boundaries # Genes in Wide Peak
7p11.2 1.1653e-12 1.1653e-12 chr7:55200240-55340844 1
1q32.1 0.024142 0.024142 chr1:203146460-205225922 32
12q14.1 0.15455 0.15455 chr12:57622487-58850360 35
Genes in Wide Peak

This is the comprehensive list of amplified genes in the wide peak for 7p11.2.

Table S1.  Genes in bold are cancer genes as defined by The Sanger Institute: Cancer Gene Census[7].

Genes
EGFR
Genes in Wide Peak

This is the comprehensive list of amplified genes in the wide peak for 1q32.1.

Table S2.  Genes in bold are cancer genes as defined by The Sanger Institute: Cancer Gene Census[7].

Genes
MDM4
ATP2B4
CHI3L1
CHIT1
FMOD
KISS1
PIK3C2B
PRELP
RBBP5
REN
SNRPE
CNTN2
BTG2
SOX13
ZC3H11A
TMCC2
LRRN2
PLEKHA6
NFASC
DSTYK
OPTC
LAX1
ETNK2
LINC00260
PPP1R15B
LOC127841
GOLT1A
LINC00303
TMEM81
SNORA77
LOC730227
ZBED6
Genes in Wide Peak

This is the comprehensive list of amplified genes in the wide peak for 12q14.1.

Table S3.  Genes in bold are cancer genes as defined by The Sanger Institute: Cancer Gene Census[7].

Genes
CDK4
DDIT3
hsa-mir-26a-2
hsa-mir-616
CYP27B1
B4GALNT1
GLI1
INHBC
KIF5A
MARS
METTL1
TSPAN31
SHMT2
TSFM
CTDSP2
DCTN2
AVIL
OS9
R3HDM2
METTL21B
NDUFA4L2
ARHGAP9
SLC26A10
PIP4K2C
INHBE
XRCC6BP1
MARCH9
MBD6
ARHGEF25
AGAP2
DTX3
STAC3
MIR26A2
LOC100130776
LOC100506844

Figure 2.  Genomic positions of deleted regions: the X-axis represents the normalized deletion signals (top) and significance by Q value (bottom). The green line represents the significance cutoff at Q value=0.25.

Table 2.  Get Full Table Deletions Table - 2 significant deletions found. Click the link in the last column to view a comprehensive list of candidate genes. If no genes were identified within the peak, the nearest gene appears in brackets.

Cytoband Q value Residual Q value Wide Peak Boundaries # Genes in Wide Peak
9p21.3 3.2558e-12 3.2558e-12 chr9:21238775-23878121 21
2q37.3 0.096154 0.096154 chr2:231497683-243199373 152
Genes in Wide Peak

This is the comprehensive list of deleted genes in the wide peak for 9p21.3.

Table S4.  Genes in bold are cancer genes as defined by The Sanger Institute: Cancer Gene Census[7].

Genes
hsa-mir-31
CDKN2A
CDKN2B
ELAVL2
IFNA1
IFNA2
IFNA5
IFNA6
IFNA8
IFNA13
IFNA14
IFNA22P
MTAP
C9orf53
KLHL9
DMRTA1
IFNE
MIR31
FLJ35282
MIR31HG
CDKN2B-AS1
Genes in Wide Peak

This is the comprehensive list of deleted genes in the wide peak for 2q37.3.

Table S5.  Genes in bold are cancer genes as defined by The Sanger Institute: Cancer Gene Census[7].

Genes
hsa-mir-3133
hsa-mir-149
hsa-mir-4269
hsa-mir-562
hsa-mir-1471
hsa-mir-1244-1
AGXT
ALPI
ALPP
ALPPL2
KIF1A
BOK
CHRND
CHRNG
COL6A3
DTYMK
GBX2
GPC1
GPR35
HDLBP
HTR2B
INPP5D
KCNJ13
NCL
NDUFA10
SEPT2
NEU2
NPPC
PDCD1
PDE6D
PPP1R7
PSMD1
PTMA
SNORD20
SAG
SPP2
DGKD
PER2
LRRFIP1
GPR55
ECEL1
EIF4E2
HDAC4
FARP2
ARL4C
RAMP1
NMUR1
STK25
COPS8
CAPN10
PASK
ATG4B
SH3BP4
NGEF
SNORD82
SNED1
GIGYF2
TRAF3IP1
ANO7
PRLH
THAP4
ANKMY1
SCLY
ASB1
CAB39
UGT1A10
UGT1A8
UGT1A7
UGT1A6
UGT1A5
UGT1A9
UGT1A4
UGT1A1
UGT1A3
ATG16L1
USP40
HJURP
HES6
CXCR7
RNPEPL1
GAL3ST2
RAB17
COPS7B
TRPM8
MLPH
IQCA1
C2orf54
ARMC9
EFHD1
ILKAP
ITM2C
ING5
MGC16025
B3GNT7
AGAP1
TWIST2
DIS3L2
NEU4
SPATA3
MTERFD2
UBE2F
OTOS
MYEOV2
OR6B3
LOC150935
LOC151171
LOC151174
LOC151475
LINC00471
LOC151484
MSL3P1
C2orf57
TIGD1
LOC200772
C2orf72
CXXC11
DUSP28
ESPNL
ECEL1P2
LOC348761
RBM44
AQP12A
KLHL30
C2orf82
OR6B2
ASB18
FLJ43879
MIR149
DNAJB3
LOC643387
PRR21
PRSS56
AQP12B
SNORA75
SCARNA6
SCARNA5
D2HGDH
LOC728323
PP14571
LOC100286922
MIR1471
MIR1244-1
BOK-AS1
MIR1244-3
MIR1244-2
MIR4269
UBE2F-SCLY
MIR4777
MIR2467
MIR4440
MIR4786
MIR4441
Arm-level results

Table 3.  Get Full Table Arm-level significance table - 12 significant results found. The significance cutoff is at Q value=0.25.

Arm # Genes Amp Frequency Amp Z score Amp Q value Del Frequency Del Z score Del Q value
1p 2121 0.07 0.373 0.962 0.22 4.89 2.53e-06
1q 1955 0.10 1.15 0.962 0.00 -1.71 0.962
2p 924 0.00 -1.71 0.962 0.02 -1.11 0.962
2q 1556 0.00 -1.77 0.962 0.00 -1.77 0.962
3p 1062 0.02 -1.13 0.962 0.00 -1.72 0.962
3q 1139 0.02 -1.11 0.962 0.02 -1.11 0.962
4p 489 0.02 -1 0.962 0.06 0.21 0.927
4q 1049 0.00 -1.54 0.962 0.21 4.97 1.9e-06
5p 270 0.00 -1.61 0.962 0.08 0.836 0.576
5q 1427 0.00 -1.69 0.962 0.08 0.642 0.694
6p 1173 0.00 -1.71 0.962 0.04 -0.53 0.962
6q 839 0.00 -1.51 0.962 0.23 5.66 5.19e-08
7p 641 0.23 5.73 1.03e-07 0.00 -1.5 0.962
7q 1277 0.42 11.5 0 0.00 -1.33 0.962
8p 580 0.04 -0.459 0.962 0.00 -1.67 0.962
8q 859 0.15 3.19 0.00943 0.00 -1.58 0.962
9p 422 0.03 -0.713 0.962 0.27 7.16 8.22e-12
9q 1113 0.02 -1.01 0.962 0.10 1.35 0.302
10p 409 0.05 -0.0185 0.962 0.26 6.63 1.68e-10
10q 1268 0.03 -0.789 0.962 0.27 6.81 6.28e-11
11p 862 0.04 -0.345 0.962 0.10 1.44 0.298
11q 1515 0.06 0.0637 0.962 0.02 -1.1 0.962
12p 575 0.06 0.232 0.962 0.04 -0.371 0.962
12q 1447 0.04 -0.475 0.962 0.06 0.108 0.962
13q 654 0.00 -1.6 0.962 0.12 2.01 0.0993
14q 1341 0.00 -1.52 0.962 0.25 6.08 4.68e-09
15q 1355 0.00 -1.67 0.962 0.10 1.26 0.321
16p 872 0.04 -0.495 0.962 0.00 -1.69 0.962
16q 702 0.04 -0.445 0.962 0.02 -1.05 0.962
17p 683 0.00 -1.69 0.962 0.02 -1.09 0.962
17q 1592 0.02 -1.18 0.962 0.00 -1.76 0.962
18p 143 0.06 0.333 0.962 0.06 0.333 0.924
18q 446 0.02 -0.998 0.962 0.06 0.216 0.927
19p 995 0.08 0.753 0.962 0.02 -1.03 0.962
19q 1709 0.12 1.5 0.665 0.38 9.44 0
20p 355 0.04 -0.403 0.962 0.02 -1.01 0.962
20q 753 0.08 0.753 0.962 0.00 -1.65 0.962
21q 509 0.00 -1.66 0.962 0.04 -0.45 0.962
22q 921 0.00 -1.64 0.962 0.10 1.34 0.302
Xq 1312 0.00 -1.74 0.962 0.02 -1.15 0.962
Methods & Data
Input
Description
  • Segmentation File: The segmentation file contains the segmented data for all the samples identified by GLAD, CBS, or some other segmentation algorithm. (See GLAD file format in the Genepattern file formats documentation.) It is a six column, tab-delimited file with an optional first line identifying the columns. Positions are in base pair units.The column headers are: (1) Sample (sample name), (2) Chromosome (chromosome number), (3) Start Position (segment start position, in bases), (4) End Position (segment end position, in bases), (5) Num markers (number of markers in segment), (6) Seg.CN (log2() -1 of copy number).

  • Markers File: The markers file identifies the marker names and positions of the markers in the original dataset (before segmentation). It is a three column, tab-delimited file with an optional header. The column headers are: (1) Marker Name, (2) Chromosome, (3) Marker Position (in bases).

  • Reference Genome: The reference genome file contains information about the location of genes and cytobands on a given build of the genome. Reference genome files are created in Matlab and are not viewable with a text editor.

  • CNV Files: There are two options for the cnv file. The first option allows CNVs to be identified by marker name. The second option allows the CNVs to be identified by genomic location. Option #1: A two column, tab-delimited file with an optional header row. The marker names given in this file must match the marker names given in the markers file. The CNV identifiers are for user use and can be arbitrary. The column headers are: (1) Marker Name, (2) CNV Identifier. Option #2: A 6 column, tab-delimited file with an optional header row. The 'CNV Identifier' is for user use and can be arbitrary. 'Narrow Region Start' and 'Narrow Region End' are also not used. The column headers are: (1) CNV Identifier, (2) Chromosome, (3) Narrow Region Start, (4) Narrow Region End, (5) Wide Region Start, (6) Wide Region End

  • Amplification Threshold: Threshold for copy number amplifications. Regions with a log2 ratio above this value are considered amplified.

  • Deletion Threshold: Threshold for copy number deletions. Regions with a log2 ratio below the negative of this value are considered deletions.

  • Cap Values: Minimum and maximum cap values on analyzed data. Regions with a log2 ratio greater than the cap are set to the cap value; regions with a log2 ratio less than -cap value are set to -cap. Values must be positive.

  • Broad Length Cutoff: Threshold used to distinguish broad from focal events, given in units of fraction of chromosome arm.

  • Remove X-Chromosome: Flag indicating whether to remove data from the X-chromosome before analysis. Allowed values= {1,0} (1: Remove X-Chromosome, 0: Do not remove X-Chromosome.

  • Confidence Level: Confidence level used to calculate the region containing a driver.

  • Join Segment Size: Smallest number of markers to allow in segments from the segmented data. Segments that contain fewer than this number of markers are joined to the neighboring segment that is closest in copy number.

  • Arm Level Peel Off: Flag set to enable arm-level peel-off of events during peak definition. The arm-level peel-off enhancement to the arbitrated peel-off method assigns all events in the same chromosome arm of the same sample to a single peak. It is useful when peaks are split by noise or chromothripsis. Allowed values= {1,0} (1: Use arm level peel off, 0: Use normal arbitrated peel-off).

  • Maximum Sample Segments: Maximum number of segments allowed for a sample in the input data. Samples with more segments than this threshold are excluded from the analysis.

  • Gene GISTIC: When enabled (value = 1), this option causes GISTIC to analyze deletions using genes instead of array markers to locate the lesion. In this mode, the copy number assigned to a gene is the lowest copy number among the markers that represent the gene.

Values

List of inputs used for this run of GISTIC2. All files listed should be included in the archived results.

  • Segmentation File = /xchip/cga/gdac-prod/tcga-gdac/jobResults/PrepareGisticDNASeq/LGG-TP/10006250/segmentationfile.txt

  • Markers File = /xchip/cga/gdac-prod/tcga-gdac/jobResults/PrepareGisticDNASeq/LGG-TP/10006250/markersfile.txt

  • Reference Genome = /xchip/cga/reference/gistic2/hg19_with_miR_20120227.mat

  • CNV Files = /xchip/cga/reference/gistic2/CNV.hg19.bypos.111213.txt

  • Amplification Threshold = 0.3

  • Deletion Threshold = 0.3

  • Cap Values = 2

  • Broad Length Cutoff = 0.5

  • Remove X-Chromosome = 0

  • Confidence Level = 0.99

  • Join Segment Size = 10

  • Arm Level Peel Off = 1

  • Maximum Sample Segments = 10000

  • Gene GISTIC = 0

Table 4.  Get Full Table First 10 out of 52 Input Tumor Samples.

Tumor Sample Names
TCGA-CS-4938-01B-11D-1891-02
TCGA-CS-4941-01A-01D-1465-02
TCGA-CS-4944-01A-01D-1465-02
TCGA-CS-5396-01A-02D-1465-02
TCGA-CS-6186-01A-12D-2022-02
TCGA-CS-6290-01A-11D-1703-02
TCGA-DB-5273-01A-01D-1465-02
TCGA-DB-5276-01A-01D-1465-02
TCGA-DB-5279-01A-01D-1465-02
TCGA-DB-5280-01A-01D-1465-02

Figure 3.  Segmented copy number profiles in the input data

Output
All Lesions File (all_lesions.conf_##.txt, where ## is the confidence level)

The all lesions file summarizes the results from the GISTIC run. It contains data about the significant regions of amplification and deletion as well as which samples are amplified or deleted in each of these regions. The identified regions are listed down the first column, and the samples are listed across the first row, starting in column 10.

Region Data

Columns 1-9 present the data about the significant regions as follows:

  1. Unique Name: A name assigned to identify the region.

  2. Descriptor: The genomic descriptor of that region.

  3. Wide Peak Limits: The 'wide peak' boundaries most likely to contain the targeted genes. These are listed in genomic coordinates and marker (or probe) indices.

  4. Peak Limits: The boundaries of the region of maximal amplification or deletion.

  5. Region Limits: The boundaries of the entire significant region of amplification or deletion.

  6. Q values: The Q value of the peak region.

  7. Residual Q values: The Q value of the peak region after removing ('peeling off') amplifications or deletions that overlap other, more significant peak regions in the same chromosome.

  8. Broad or Focal: Identifies whether the region reaches significance due primarily to broad events (called 'broad'), focal events (called 'focal'), or independently significant broad and focal events (called 'both').

  9. Amplitude Threshold: Key giving the meaning of values in the subsequent columns associated with each sample.

Sample Data

Each of the analyzed samples is represented in one of the columns following the lesion data (columns 10 through end). The data contained in these columns varies slightly by section of the file. The first section can be identified by the key given in column 9 - it starts in row 2 and continues until the row that reads 'Actual Copy Change Given.' This section contains summarized data for each sample. A '0' indicates that the copy number of the sample was not amplified or deleted beyond the threshold amount in that peak region. A '1' indicates that the sample had low-level copy number aberrations (exceeding the low threshold indicated in column 9), and a '2' indicates that the sample had high-level copy number aberrations (exceeding the high threshold indicated in column 9).The second section can be identified the rows in which column 9 reads 'Actual Copy Change Given.' The second section exactly reproduces the first section, except that here the actual changes in copy number are provided rather than zeroes, ones, and twos.The final section is similar to the first section, except that here only broad events are included. A 1 in the samples columns (columns 10+) indicates that the median copy number of the sample across the entire significant region exceeded the threshold given in column 9. That is, it indicates whether the sample had a geographically extended event, rather than a focal amplification or deletion covering little more than the peak region.

Amplification Genes File (amp_genes.conf_##.txt, where ## is the confidence level)

The amp genes file contains one column for each amplification peak identified in the GISTIC analysis. The first four rows are:

  1. Cytoband

  2. Q value

  3. Residual Q value

  4. Wide Peak Boundaries

These rows identify the lesion in the same way as the all lesions file.The remaining rows list the genes contained in each wide peak. For peaks that contain no genes, the nearest gene is listed in brackets.

Deletion Genes File (del_genes.conf_##.txt, where ## is the confidence level)

The del genes file contains one column for each deletion peak identified in the GISTIC analysis. The file format for the del genes file is identical to the format for the amp genes file.

Gistic Scores File (scores.gistic)

The scores file lists the Q values [presented as -log10(q)], G scores, average amplitudes among aberrant samples, and frequency of aberration, across the genome for both amplifications and deletions. The scores file is viewable with the Genepattern SNPViewer module and may be imported into the Integrated Genomics Viewer (IGV).

Segmented Copy Number (raw_copy_number.{fig|pdf|png} )

The segmented copy number is a pdf file containing a colormap image of the segmented copy number profiles in the input data.

Amplification Score GISTIC plot (amp_qplot.{fig|pdf|png|v2.pdf})

The amplification pdf is a plot of the G scores (top) and Q values (bottom) with respect to amplifications for all markers over the entire region analyzed.

Deletion Score GISTIC plot (del_qplot.{fig|pdf|png|v2.pdf})

The deletion pdf is a plot of the G scores (top) and Q values (bottom) with respect to deletions for all markers over the entire region analyzed.

Tables (table_{amp|del}.conf_##.txt, where ## is the confidence level)

Tables of basic information about the genomic regions (peaks) that GISTIC determined to be significantly amplified or deleted. These describe three kinds of peak boundaries, and list the genes contained in two of them. The region start and region end columns (along with the chromosome column) delimit the entire area containing the peak that is above the significance level. The region may be the same for multiple peaks. The peak start and end delimit the maximum value of the peak. The extended peak is the peak determined by robust, and is contained within the wide peak reported in {amp|del}_genes.txt by one marker.

Broad Significance Results (broad_significance_results.txt)

A table of per-arm statistical results for the data set. Each arm is a row in the table. The first column specifies the arm and the second column counts the number of genes known to be on the arm. For both amplification and deletion, the table has columns for the frequency of amplification or deletion of the arm, and a Z score and Q value.

Broad Values By Arm (broad_values_by_arm.txt)

A table of chromosome arm amplification levels for each sample. Each row is a chromosome arm, and each column a sample. The data are in units of absolute copy number -2.

All Data By Genes (all_data_by_genes.txt)

A gene-level table of copy number values for all samples. Each row is the data for a gene. The first three columns name the gene, its NIH locus ID, and its cytoband - the remaining columns are the samples. The copy number values in the table are in units of (copy number -2), so that no amplification or deletion is 0, genes with amplifications have positive values, and genes with deletions are negative values. The data are converted from marker level to gene level using the extreme method: a gene is assigned the greatest amplification or the least deletion value among the markers it covers.

Broad Data By Genes (broad_data_by_genes.txt)

A gene-level table of copy number data similar to the all_data_by_genes.txt output, but using only broad events with lengths greater than the broad length cutoff. The structure of the file and the methods and units used for the data analysis are otherwise identical to all_data_by_genes.txt.

Focal Data By Genes (focal_data_by_genes.txt)

A gene-level table of copy number data similar to the all_data_by_genes.txt output, but using only focal events with lengths greater than the focal length cutoff. The structure of the file and the methods and units used for the data analysis are otherwise identical to all_data_by_genes.txt.

All Thresholded By Genes (all_thresholded.by_genes.txt)

A gene-level table of discrete amplification and deletion indicators at for all samples. There is a row for each gene. The first three columns name the gene, its NIH locus ID, and its cytoband - the remaining columns are the samples. A table value of 0 means no amplification or deletion above the threshold. Amplifications are positive numbers: 1 means amplification above the amplification threshold; 2 means amplifications larger to the arm level amplifications observed for the sample. Deletions are represented by negative table values: -1 represents deletion beyond the threshold; -2 means deletions greater than the minimum arm-level deletion observed for the sample.

Sample Cutoffs (sample_cutoffs.txt)

A table of the per-sample threshold cutoffs (in units of absolute copy number -2) used to distinguish the high level amplifications (+/-2) from ordinary amplifications (+/-1) in the all_thresholded.by_genes.txt output file. The table contains three columns: the sample identifier followed by the low (deletion) and high (amplification) cutoff values. The cutoffs are calculated as the minimum arm-level amplification level less the deletion threshold for deletions and the maximum arm-level amplification plus the amplification threshold for amplifications.

Focal Input To Gistic (focal_input.seg.txt)

A list of copy number segments describing just the focal events present in the data. The segment amplification/deletion levels are in units of (copy number -2), with amplifications positive and deletions negative numbers. This file may be viewed with IGV.

Gene Counts vs. Copy Number Alteration Frequency (freqarms_vs_ngenes.{fig|pdf})

An image showing the correlation between gene counts and frequency of copy number alterations.

Confidence Intervals (regions_track.conf_##.bed, where ## is the confidence level)

A file indicating the position of the confidence intervals around GISTIC peaks that can be loaded as a track in a compatible viewer browser such as IGV or the UCSC genome browser.

GISTIC

GISTIC identifies genomic regions that are significantly gained or lost across a set of tumors. It takes segmented copy number ratios as input, separates arm-level events from focal events, and then performs two tests: (i) identifies significantly amplified/deleted chromosome arms; and (ii) identifies regions that are significantly focally amplified or deleted. For the focal analysis, the significance levels (Q values) are calculated by comparing the observed gains/losses at each locus to those obtained by randomly permuting the events along the genome to reflect the null hypothesis that they are all 'passengers' and could have occurred anywhere. The locus-specific significance levels are then corrected for multiple hypothesis testing. The arm-level significance is calculated by comparing the frequency of gains/losses of each arm to the expected rate given its size. The method outputs genomic views of significantly amplified and deleted regions, as well as a table of genes with gain or loss scores. A more in depth discussion of the GISTIC algorithm and its utility is given in [1], [3], and [5].

CNV Description

Regions of the genome that are prone to germ line variations in copy number are excluded from the GISTIC analysis using a list of germ line copy number variations (CNVs). A CNV is a DNA sequence that may be found at different copy numbers in the germ line of two different individuals. Such germ line variations can confound a GISTIC analysis, which finds significant somatic copy number variations in cancer. A more in depth discussion is provided in [6]. GISTIC currently uses two CNV exclusion lists. One is based on the literature describing copy number variation, and a second one comes from an analysis of significant variations among the blood normals in the TCGA data set.

Download Results

In addition to the links below, the full results of the analysis summarized in this report can also be downloaded programmatically using firehose_get, or interactively from either the Broad GDAC website or TCGA Data Coordination Center Portal.

References
[1] Beroukhim et al, Assessing the significance of chromosomal aberrations in cancer: Methodology and application to glioma, Proc Natl Acad Sci U S A. Vol. 104:50 (2007)
[3] Mermel et al, GISTIC2.0 facilitates sensitive and confident localization of the targets of focal somatic copy-number alteration in human cancers, Genome Biology Vol. 12:4 (2011)
[5] Beroukhim et al., The landscape of somatic copy-number alteration across human cancers, Nature Vol. 463:7283 (2010)
[6] McCarroll, S. A. et al., Integrated detection and population-genetic analysis of SNPs and copy number variation, Nat Genet Vol. 40(10):1166-1174 (2008)