Chromosomal microarray (CMA) analysis to detect copy number variations (CNVs) —duplication or deletion of chromosomal regions — is the recommended first-tier diagnostic screen for autism, and studies of structural variation (SV) to date have largely focused on CNV detection and interpretation. Copy-neutral SVs, or balanced chromosomal abnormalities (BCAs), have not been systematically surveyed in autism, suggesting that these rearrangements could represent a potent class of loss-of-function mutations that have not yet been characterized in autism.

In this study, Michael Talkowski and his colleagues performed whole-genome sequencing using large insert jumping libraries to delineate the full mutational spectrum of SVs in individuals with autism in a subset of the Simons Simplex Collection (259 families). The jumping library strategy is a unique approach that sequences millions of fragments of DNA that have been manipulated so that the ends of the fragment are separated by several thousand bases in the genome. They also developed an SV sequencing classifier algorithm, and identified all CNVs and BCAs in autism at a resolution of 5-10 kb. Their analyses provided strong concordance with large CNV calls detectable by CMA, and revealed myriad complex SVs. One type of SV that they detected is a remarkably common class of variation involving inversion flanked by two duplications at the breakpoint (dupINVdup) that was not previously described in population-based SV studies.

This ‘paired-duplication’ pattern was the most common hallmark for a complex class of duplication-associated SVs that were detected in 12.4 percent of the individuals with autism evaluated in this study. Overall, Talkowski and his group found that 5.6 percent of all duplications detected by CMA in this cohort were misclassified and actually represent complex cryptic SVs. They found that at least 80 percent of loss-of-function mutations that result from SVs were cryptic to detection by CMA, and that whole-genome sequencing at approximately 5-10 kb resolution led to an approximately twofold increase in potential pathogenic SVs in this cohort.

The results of this pilot study suggest that whole-genome sequencing in the entire Simons Simplex Collection cohort is likely to yield a significant burden of cryptic SVs that contribute to autism etiology, and that comprehensive evaluation of complex SVs such as the ‘paired duplications’ and dupINVdup variations detected in this study is warranted to determine their impact in autism. Talkowski and his team’s results also suggest that rare ‘paired-duplication’ patterns detected by CMA in diagnostic practice may warrant further scrutiny by high-resolution methods.