Investigation of patients with mental retardation and dysmorphic features using comparative genomic hybridization and subtelomeric multiplex ligation dependent probe amplication

For a number of years we have used high resolution metaphase comparative genomic hybridization (CGH, referred to as HR-CGH in our prior papers) for detection of cryptic chromosomal imbalances in patients with mental retardation and dysmorphic features. Although this technique detects imbalances in as many as 12% of patients, additional diagnoses can be obtained if CGH is combined with subtelomeric FISH analysis [Kirchhoff et al., 2004]. In this study, we set out to test the diagnostic yield of subtelomeric multiplex ligation-dependent probe amplification (MLPA) in patients already investigated by CGH. In this study, we present the CGH and subtelomeric MLPA results of 258 patients with mental retardation and dysmorphic features. Introduction of a newmethod in a diagnostic routine setting requires careful consideration regarding evaluation of results. Thus, the diagnostic criteria for the MLPA analysis and the strategy used for exclusion of putative polymorphisms are discussed. The MLPA assay was performed using the SALSA P019, P020, and P036 human subtelomeric probe sets (MRC-Holland, Amsterdam, The Netherlands). Samples of DNA from 20 normal individuals were used to create reference values for eachMLPAprobe set. The references for P019, P020, and P036 werebased on86, 81, and92analyses respectively.Henceforth, P019 and P020 aremerged and referred to as a single probe set as combination of these two sets interrogate all subtelomeric regions (except the acrocentric p-arms) as does the P036 probe set. Following capillary electrophoresis on an ABI Prism 3100, peak areas were normalized in relation to the mean of neighboring peaks. The ratio of each peak and the correspondingmean peak area of the reference were computed, and it was measured in SDs how far each peak area was from the corresponding mean area of the reference. The ‘‘area distance in SDs’’ reflects the individual probe reliability, because a peak ratio indicating an imbalance get a low area distance in SDs, if the probe shows extensive variation (i.e., high SD value) in the reference data set; hereby false positives may be spotted (for normally distributed data the probability for observing an area outside 3 SD is less than 1%) [Gerdes et al., 2005; Gerdes et al., in press] (details of data analysis and downloads of the software are available on www.chromosomelab.dk). Thirteen cases with 20 known aberrations in subtelomeric regions (12 deletions and eight duplications) were used to set up a diagnostic threshold, which was based on both peak area ratios and the area distance in SDs. The average ratio of the deletion probes to the reference was 0.50 with a range of 0.40– 0.60. The average area distance in SDs was 7.4 with a range of 4.1–11.8. The average ratio of the duplication probes to the referencewas 1.46with a range of 1.35–1.64. The average area distance in SDs was 7.7 with a range of 4.4–10.2. On the basis of these values, a diagnostic threshold was set: a peak was consideredabnormalwhen the peakareawasmore than4.0SDs from the reference mean and the ratio was 1.35. The diagnostic threshold was tested on samples from 258 consecutive patients with mental retardation and dysmorphic features who were referred for CGH analysis. The results are summarized in Table I. Abnormalities were detected in 29 (11.2%) patients by subtelomericMLPA. In 14 of those 29 cases (5.4% of 258), the imbalances were detected by both probe sets (in cases 8, 11, and 29 one of the ratio values or one of the area distances in SDs values were close to, but did not reach, the diagnostic threshold) and in remaining 15 cases (5.8% of 258) the imbalances were detected by one probe set. In two cases, the P019/P020 probe set showed a deletion of 3pter. Sequencing analysis showed that these patients had aG instead of an A in the splice-site of the probe. This probe is now recognized to detect a polymorphic target sequence (J. Schouten, MRC-Holland, personal communication), and has been replaced in a new generation of subtelomeric probe sets (www. MRC-Holland.com). Thus, the two 3pter deletions are considered to be false positive results caused by a single nucleotide polymorphism (SNP). Recent studies have shown that there is considerable structural variation in the genomes of normal individuals [Iafrate et al., 2004; Sebat et al., 2004]. In the design of subtelomeric MLPA probe sets, a compromise is made between avoiding polymorphisms and selecting target sequences close to the telomeres, in order tomaximize the sensitivity of the screen. In order to reduce the risk of detecting polymorphisms, imbalances detected by only a single probe set were assessed as clinically insignificant. We made this decision for two reasons. First,with regard to small duplications/deletions there is a risk that the clinical significance remains unresolved or a false diagnosis is made. This is distressing for both the family and the medical professionals and genetic counseling becomes inadequate. Second, when submicroscopic imbalances are detected, follow-up is often labor-intensive and costly. In the present study, 15 duplications were detected by one probe set only. Apart from two duplications, all were detected in two, three, or five patients, which is consistent with the assumption that they represent polymorphisms. Indeed, a duplication of *Correspondence to: Dr. Maria Kirchhoff, Rigshospitalet, Department of Clinical Genetics, 4052, Blegdamsvej 9, Copenhagen, Denmark DK-2100. E-mail: markir@rh.dk