Abstract

Background: Using array techniques, it was recently shown that about 10% of patients with mental retardation of unknown origin harbour cryptic chromosomal aneusomies. However, data analysis is currently not standardised and little is known about its sensitivity and specificity.

Methods: We have developed an electronic data analysis tool for gene-mapping SNP arrays, a software tool that we call Copy Number Variation Finder (CNVF). Using CNVF, we analysed 104 unselected patients with mental retardation of unknown origin with a genechip mapping 100K SNP array and established an optimised set of analysis parameters.

Results: We detected deletions as small as 20 kb when covered by at least three single-nucleotide polymorphisms (SNPs) and duplications as small as 150 kb when covered by at least six SNPs, with only one false-positive signal in six patients. In 9.1% of patients, we detected apparently disease-causing or de novo aberrations ranging in size from 0.4 to 14 Mb. Morphological anomalies in patients with de novo aberrations were equal to that of unselected patients when measured with de Vries score.

Conclusion: Our standardised CNVF data analysis tool is easy to use and has high sensitivity and specificity. As some genomic regions are covered more densely than others, the genome-wide resolution of the 100K array is about 400–500 kb for deletions and 900–1000 kb for duplications. The detection rate of about 10% of de novo aberrations is independent of selection of patients for particular features. The incidental finding in two patients of heterozygosity for the 250 kb recurrent deletion at the NPH1 locus, associated with autosomal recessive juvenile nephronophthisis, which was inherited from a healthy parent, highlights the fact that inherited aberrations might be disease-related even though not causal for mental retardation.