Identification of pathogenic CNVs in a simplex autism cohort and measurement of effect size on cognitive, adaptive, and social function. A. E. Hare1, D. Moreno De Luca2, K. B. Boomer3, S. J. Sanders4,5, M. W. State4,5, M. Benedetti6, A. L. Beaudet7, E. H. Cook8, D. M. Martin9, D. H. Ledbetter1, C. L. Martin1 1) Autism & Developmental Medicine Institute, Geisinger Health System, Danville, PA; 2) Department of Psychiatry, Yale University, New Haven, CT; 3) Department of Mathematics, Bucknell University, Lewisburg, PA; 4) Department of Genetics, Yale University, New Haven, CT; 5) Department of Psychiatry, UCSF, San Francisco, CA; 6) Simons Foundation, New York, NY; 7) Department of Human and Molecular Genetics, Baylor College of Medicine, Houston, TX; 8) Institute for Juvenile Research, Department of Psychiatry, University of Illinois, Chicago, IL; 9) Departments of Pediatrics and Human Genetics, University of Michigan Medical Center, Ann Arbor, MI.

   We conducted genome-wide copy number variant (CNV) analysis in 2,742 autism probands from the Simons Simplex Collection (SSC). CNVs 250 kb in size were classified as pathogenic based on clinical interpretation guidelines as determined by an expert panel of clinical cytogeneticists and geneticists. We analyzed de novo and inherited cases of recurrent (rec) CNVs (mediated by flanking segmental duplications) and de novo cases of non-recurrent (nrec) CNVs. This approach allowed us to identify rare, clinically significant pCNVs, which would not have reached statistical significance in a case/control study, by incorporating previously established evidence for pathogenicity. While statistically significant CNVs were previously identified in 1% of SSC cases, we identified pCNVs in 4.3% of cases, including 55 deletions (del), 56 duplications (dup), 2 unbalanced translocations and 6 sex chromosome aneuploidies (1 XXX, 2 XXY, and 3 XYY; removed from phenotypic analyses). pCNVs were identified in 13 rec and 41 nrec genomic regions. The most common rec pCNVs were dup 16p11.2 (n=12), dup 1q21.1 (n=11) and del 16p11.2 (n=9); all nrec pCNVs were unique. We measured the effect of carrying a pCNV on cognitive, adaptive, and social function by dividing the SSC probands into 5 categories: 1) no pCNV (nonCNV); 2) de novo del (dn-del); 3) de novo dup (dn-dup); 4) inherited del (inh-del); and 5) inherited dup (inh-dup). Dn-del probands had significantly lower cognitive (p0.001) and adaptive (p0.001) function than nonCNV probands. Cognitive ability of dn-del probands was influenced by pCNV size (-2.66 IQ points/Mb) and number of genes per pCNV (-0.43 IQ points/gene). Not surprisingly, dn-del probands did not differ from nonCNV probands in social function since the SSC cohort was ascertained by meeting full criteria for an autism diagnosis. However, we hypothesized that the presence of a pCNV may have a significant effect on a probands Social Responsiveness Scale (SRS) raw score relative to their unaffected family members. We observed a 3.9 standard deviation (SD) shift between dn-del probands and their parents SRS scores that was significantly higher (p=0.033) than the 3.4 SD shift observed in nonCNV probands. Dn-dup, inh-del, or inh-dup probands did not differ from nonCNV probands on any phenotypic measures, including familial SRS shift. These findings demonstrate that dn-del probands have a more deleterious phenotypic effect on cognitive, adaptive, and social function.

You may contact the first author (during and after the meeting) at