Analysis of aberrant and normal copy-number variation of human chromosome 22

Abstract

Copy-Number Variation (CNV) and Copy-Number Polymorphisms (CNP) have emerged as pervasive architectural feature of the human genome and are expected to contribute significantly to phenotypic variation both in the healthy individual and in disease states. Array-CGH, the dominant methodology to determine CNV, has a typical predictive resolution of about 50 kbp. CNV below that horizon will he missed as will be the actual breakpoint-sequence of the variant or aberration. For a complete catalog of all sizes-classes of CNV, a better understanding of the mechanisms of breakage and the detection of possible fusion genes or exon-copy-number variation, a procedure that predicts CNV at sub-exon level and follows up with sequencing across the breakpoints is needed. We have developed HR-CGH based on high-density oligonucleotide tiling microarrays [Urban, Korbel et al. PNAS 2006; Korbel, Urban et al. PNAS 2007]. Maskless Synthesis is used to generate arrays with 385 000 oligomers that are designed to represent the non-repetitive part of the genomic sequence of human chromosome 22 at a tiling density of 1 oligomer/85bp. The arrays are probed with fragmented full-complexity genomic DNA that has been fluorescently labeled. The ratio of signal intensities from the control and experimental channels are processed by the BreakPtr algorithm, which predicts copy-number changes, their dosages and breakpoints while screening out false-positive calls caused by cross-hybridization. Predictions are validated with vectorette-PCR and direct sequencing of the resulting amplicons, making cloning superfluous. Using this approach we have studied various types of disease causing aberrations on chromosome 22 as well as CNVs and CNPs, and their breakpoints. We probed the microarrays with samples from probands with Velo-Cardio-Facial Syndrome (VCFS, 22q11DS), Cat-Eye-Syndrome (CES), Emanuel-Syndrome, Duplication22-Syndrome and from the HapMap panel. We have detected, cataloged and verified variations from 3 Mb to smaller than 1kb in size and assigned breakpoint coordinates to them that we expect to be generally accurate to within 1 kbp of sequence or better. In several cases we were already successful in determining the exact breakpoint sequence. We expect that HR-CGH will be a method of broad and general utility in the study of aberrations and variations on chromosome 22 but also across the entire human genome.