Genomic contributions to self-injurious behavior in autism spectrum disorder

Self-injurious behavior is a serious behavioral disorder characterized by the infliction of physical injury to the individual’s own body, resulting in tissue damage ranging from minor to permanent. In its most severe form SIB can have potentially life-threatening implica¬tions. Although a high prevalence of SIB has been reported in genetic intellectual disability syndromes such as Lesch-Nyhan Syndrome, Smith-Magenis Syndrome, and Cornelia de Lange Syndrome, studies of the genomic landscape of SIB are limited. The mutation of a specific gene and/or region resulting in a distinct type of SIB as in these genetic ID syndromes provides evidence of genetic predisposition to SIB. Only a few studies have focused on elucidating the role of genes in SIB, with much of prior literature focused on the behavioral assessment and treatment of SIB. To fully understand the breadth of genetic susceptibility to SIB, we assess SIB among two datasets: (1) 34 trios of unaffected parents and probands with severe SIB; (2) MSSNG, an ASD resource comprising of genotypic and phenotypic data on ~11,000 indi¬viduals (~5,000 affected). The rich genomic findings in first cohort include candidate de novo single nucleotide variants, small insertions and deletions, and copy number variants of significance. These include de novo loss of function mutations in CSDE1, ANKRD11 and TRIP12 as well as a tandem duplication of the ZEB2 gene, heterozygous deletion of the 16p11.2 region, and a three-way unbalanced translocation causing 3q28qter duplication and 10q26.2qter deletion. Analysis of the MSSNG cohort required the use Google’s BigQuery to query the available whole genome sequencing and phenotype data for further statistical comparisons. We found a statistically significant association between the presence of SIB and lower IQ scores (IQ Leiter test, p = 1.74e-05) as well as statistically significant associations between SIB and the presence of de novo loss of function variants within known self-injurious behavior genes (p < 0.02) and within neurodevelopment disorder genes (p = 0.03), with or without adjustment for possible confounder, relatedness and multiple test correction. Together, these findings suggest that genomic analyses of SIB may contribute to a more comprehensive understanding of its phenotypic implications among affected individuals