Improved detection of global copy number variation using high density, non-polymorphic oligonucleotide probes

Fan Shen¹ , Jing Huang¹ , Karen R Fitch¹ , Vivi B Truong¹ , Andrew Kirby² , Wenwei Chen¹ , Jane Zhang¹ , Guoying Liu¹ , Steven A McCarroll³ , Keith W Jones¹ and Michael H Shapero¹

¹Affymetrix, Inc. 3420 Central Expressway; Santa Clara, CA 95051, USA

²Center for Human Genetic Research, Massachusetts General Hospital, Boston, MA 02114, USA

³Program in Medical and Population Genetics, Broad Institute of MIT and Harvard, Cambridge, Massachusetts 02142, USA

author email corresponding author email

BMC Genetics 2008, 9:27doi:10.1186/1471-2156-9-27


Published:	28 March 2008

Additional files

Additional file 1:

Dose response plots of a representative 1X–5X data set. Panels a-d show the scatter plots of standardized natural log intensity of the 1X, 3X, 4X, and 5X samples relative to the 2X sample. Here, standardization refers to the following data transformation: standardized intensity of chromosome X probe = (intensity of chromosome X probe-mean intensity of the autosomal probes)/standard deviation of the intensity of autosomal probes. Red dots represent randomly selected chromosome X probes and black dots represent randomly selected autosomal probes. The blues lines are the Y = X lines. Panel e shows the relationship between the natural log-transformed intensity and the natural log-transformed copy number. Natural log-transformed mean intensity of all chromosome X probes from the 1X–5X samples are plotted on the Y-axis and natural log-transformed copy number are plotted on the X-axis. The blue line is the linear regression line using the natural log-transformed mean intensity as response and natural log-transformed copy number as predictors.

Format: PDF Size: 244KB Download file

This file can be viewed with: Adobe Acrobat Reader

Additional file 2:

Dose response of probes deteriorates as the number of genomic hits increases. Panel a shows the frequency distribution of genomic matches for a set of 80,000 randomly selected chromosome X probes. Panels b-c are box-plots showing the distribution of linear correlation coefficient and regression slope grouped by the number of genomic hits of a set of 80000 randomly selected chromosome X probes. Panel d shows chromosome X hits frequency distribution of the same set of randomly selected 80000 chromosome X probes. Panels e-f are box-plots showing the distribution of linear correlation coefficient and regression slope grouped by the number of chromosome X hits of this set of 80,000 randomly selected chromosome X probes. Natural log-transformed normalized (as described in Methods) intensity of chromosome X probes of a representative set of 1X–5X samples and natural log-transformed copy number were used to calculate linear correlation coefficient and regression slope for each probe.

Format: JPEG Size: 148KB Download file

Additional file 3:

A 2-dimensional histogram showing the distribution of regression slope along with the distribution of natural log-transformed intensity. Natural log-transformed normalized (as described in Methods) intensity of 80,000 randomly selected chromosome X probes of a representative set of 1X–5X samples and natural log-transformed copy number were used to calculate the regression slope. The black vertical line denotes the maximum log intensity ratio and the green vertical line denotes the top 8% log intensity, above which there are few probes with high regression slopes. The top 10% intensity is used as the cut-off threshold in the probe filtering process.

Format: JPEG Size: 111KB Download file

Additional file 4:

Number of remaining probes and fragments following probe filtering for 3 replicates of 1X–5X samples. The data indicates the number of probes and fragments that have been retained after probe filtering for 3 replicates of the 1X–5X DNA samples.

Format: XLS Size: 14KB Download file

This file can be viewed with: Microsoft Excel Viewer

Additional file 5:

Dose response of probes improves after probe filtering and RMA procedure. Natural log-transformed normalized (as described in Methods) intensity of 80,000 randomly selected chromosome X probes of a representative set of 1X–5X DNA samples and natural log-transformed copy number were used to calculate linear correlation coefficient and regression slope for all probes(blue bars), natural log-transformed normalized intensity of post-filtering 64,035 of the 80,000 randomly selected chromosome X probes and natural log-transformed copy number were used to calculate linear correlation coefficient and regression slope for the filtered probes(grey bars), and natural log-transformed post-RMA chromosome X probe set intensity and natural log-transformed copy number were used to calculate linear correlation coefficient and regression slope for the fragments (red bars).

Format: JPEG Size: 119KB Download file

Additional file 6:

List of QPCR data and CNV coordinates. Table A represents the coordinates of CNVs in NA15510 vs. NA10851. Table B summarizes QPCR results for NA15510 vs. NA10851. Table C represents QPCR results for the CNV border analysis. Table D represents QPCR results for Mendelian inheritance (MI) errors. Table E lists counts of CNVs in HapMap trio samples NA10846-NA12144-NA12125 and NA10831-NA12155-NA12156.

Format: XLS Size: 178KB Download file

This file can be viewed with: Microsoft Excel Viewer

Additional file 7:

Chromosome X QPCR Analysis. The data represents QPCR analysis of seven independent X chromosome amplicons that were each analyzed using the 1X to 5X DNA samples.

Format: XLS Size: 17KB Download file

This file can be viewed with: Microsoft Excel Viewer