Impact of compound heterozygous events involving deletions and sequence-level variants in autism

The genomic architecture of autism is complex, with contributions from many types of genetic variation. While most autism-associated rare variants appear to act via dominant inheritance patterns, recessive events have also been implicated in autism susceptibility. However, previous studies of autism-associated recessive variation have focused only on events where both alleles are affected by sequence-level variants (single nucleotide variants and indels). Here, we focus on compound heterozygous events involving a large (>50 bp) exonic deletion on one allele and a functional sequence-level variant (loss-of-function or missense) on the other allele, which we term deletion compound heterozygous (DelCH) events. We explored four methods for assessing the impact of DelCH events on autism susceptibility. First, we compared the burden of sequence-level variants within rare exonic deletions between probands and the deletion-transmitting parent, with the hypothesis that probands will have a higher burden. Second, we compared the burden of DelCH events in probands versus their family members, i.e., unaffected siblings and parents. Third, we used a transmission disequilibrium test to evaluate the hypothesis that functional sequence-level variants within deletion-impacted genes are over-transmitted from parents to probands. Finally, we evaluated the hypothesis that autistic children have a higher density of functional sequence-level variants within deleted regions than their unaffected siblings. We applied all four methods to sequence-level variants and deletions detected from whole-genome sequencing data from autistic individuals and their family members in the MSSNG, Simons Simplex Collection (SSC), and Simons Foundation Powering Autism Research (SPARK) cohorts. Findings from all four methods were suggestive of a role for DelCH events in autism susceptibility; however, the ability to obtain statistical significance was hampered by the rarity of DelCH events and insufficient sample sizes. As autism whole-genome sequencing cohorts continue to grow, we anticipate that the methods presented here will allow researchers to re-assess the role of DelCH events in autism with greater statistical power.