1. ¹CSS Hospital, IRCSS, San Giovanni Rotondo and CSS-Mendel Institute, Rome, Italy
2. ²Department of Cancer Biology, Kimmel Cancer Center, Thomas Jefferson University, Jefferson Medical College, Philadelphia, PA, USA
3. ³Stella Maris Clinical Research Institute for Child and Adolescent Neurology and Psychiatry, Pisa, Italy
4. ⁴Medical Genetics, IRCCS Bambino Gesù Pediatric Hospital, Rome, Italy
5. ⁵Department of Pediatrics, Cattolica University, Rome, Italy
6. ⁶Department of Experimental Medicine, Sapienza University, Rome, Italy
7. ⁷Cardeza Foundation for Hematologic Research, Department of Medicine, Jefferson Medical College, Thomas Jefferson University, Philadelphia, PA, USA

Correspondence: Dr P Fortina, Kimmel Cancer Center, Thomas Jefferson University, 233 South 10th Street, 1009-J BLSB, Philadelphia, PA 19107, USA. Tel: +1 215 955 0683; Fax: +1 215 503 9142; E-mail: paolo.fortina@jefferson.edu

Received 22 April 2009; Revised 11 August 2009; Accepted 28 August 2009; Published online 7 October 2009.

Abstract

We used Affymetrix 6.0 GeneChip SNP arrays to characterize copy number variations (CNVs) in a cohort of 70 patients previously characterized on lower-density oligonucleotide arrays affected by idiopathic mental retardation and dysmorphic features. The SNP array platform includes 900 000 SNP probes and 900 000 non-SNP oligonucleotide probes at an average distance of 0.7 Kb, which facilitates coverage of the whole genome, including coding and noncoding regions. The high density of probes is critical for detecting small CNVs, but it can lead to data interpretation problems. To reduce the number of false positives, parameters were set to consider only imbalances >75 Kb encompassing at least 80 probe sets. The higher resolution of the SNP array platform confirmed the increased ability to detect small CNVs, although more than 80% of these CNVs overlapped to copy number 'neutral' polymorphism regions and 4.4% of them did not contain known genes. In our cohort of 70 patients, of the 51 previously evaluated as 'normal' on the Agilent 44K array, the SNP array platform disclosed six additional CNV changes, including three in three patients, which may be pathogenic. This suggests that about 6% of individuals classified as 'normal' using the lower-density oligonucleotide array could be found to be affected by a genomic disorder when evaluated with the higher-density microarray platforms.