Whole-genome arrays are an effective tool for identifying copy number variants (CNVs) and integral genes contributing to autism and/or intellectual disability (ASD/ID). Whilst higher resolution arrays can identify smaller CNVs, their confirmation requires molecular methods and their clinical relevance is often ambiguous. Reporting thresholds of 200kb for deletion and 500kb for duplication is common in clinical cytogenetic laboratories.
Objectives:
Our objective is to test the utility of the Affymetrix® Cytogenetics Whole-Genome 2.7M Array (Cyto2.7M Array) for detecting smaller, sub-threshold CNVs of potential clinical relevance.
Methods:
We applied Affymetrix Cyto2.7M Array for validating 39 positive CNVs previously identified by lower resolution arrays in 30 subjects with ASD and/or ID. Then, we screened a further 52 subject with ASD/ID using this platform. A total of 22 small, unique CNVs containing genes of potential ASD relevance but not previously detected by lower resolution arrays and/or under the clinical threshold were selected for investigation by FISH and/or Quantitative Multiplex PCR of Short fluorescent Fragments (QMPSF).
Results:
All of the 39 positive CNVs previously identified were confirmed using the Cyto 2.7M Array. Four out of ten small, unique CNVs (100-400Kb) previously undetected by lower resolution arrays were confirmed by FISH and/or QMPSF (40%). A further screen of 52 new subjects with ASD+/-ID using this array platform uncovered 10 unique CNVs above the clinical threshold with 5 considered pathogenic, involving 4p14, 12q24.31, 14q32.31, 15q13, and 17p13.3. Twelve small unique CNVs below the clinical threshold (49~480 Kb) containing neurofunctional genes were selected and 9/12 of them were confirmed by QMPSF (75%); 8 were familial and 1 de novo. Putatively pathogenic CNVs included: a maternally transmitted duplication (130 Kb) spanning exons 64-79 of the DMD gene which was found in a 3-year old boy manifesting autism and mild neuromotor delays.
Conclusions:
Our study established that quality control criteria based on low waviness-segment-count decreased the number of false positive small CNVs from 60% to 25%. The study also identified small CNVs of putative pathogenic importance that could be missed using current standard clinical array reporting thresholds.