Epidemiology of Consanguineous Families in Autism

Background: Analyses of large autism datasets have provided statistical and functional evidence for the role of rare point mutations and transmitted and de novo copy number variants (CNVs), and offer crucial insights into the diverse genetic mechanisms that can lead to Autism Spectrum Disorders (ASDs). 

Objectives: Ascertainment can point to differences in the underlying genetic mechanisms that lead to Autism. We subdivide families using two categories based on number of affected individuals and overall homozygosity.  The two categories serve as a remarkably effective proxy for the underlying genetics.  While the number of affected individuals, siblings as well as first-degree relatives, correlates strongly with severity of phenotype and M/F ratio, the level of homozygosity distinguishes between inherited and sporadic ASD.

Methods: We present homozygosity-CNV analysis for a cohort of 183 consanguineous families with one or more children affected with ASD 9 (HMCA). We follow up findings and draw comparisons with additional large ASD and control datasets: the Simons Simplex (SSC) and the Autism Genetic Resource Exchange (AGRE) collections (2,670 affected individuals; 9681 total individuals).

Results: We observe excess homozygosity among affected individuals relative to their unaffected siblings.  The same trend is present in two additional, independent samples: the SSC and AGRE collections enriched respectively for sporadic and inherited causes of ASD. We apply a formal test based on family quartets to evaluate excess homozygosity showing the same results. A plausible genetic mechanism driving excess homozygosity among affected individuals is the presence of homozygous copy number variation.  We show for the first time conclusive evidence that double deletions contribute significantly to ASD. 

Conclusions: Specifically, we find that de novo CNVs play a significant role in non-consanguineous families with a single affected child (p=0.04), but a lesser role in multiplex families, and they are no more common in ASD cases than controls in multiplex consanguineous families. In contrast, we present the strongest statistical evidence (p=0.013) to date that homozygous deletions, are a major contributor to ASD disease burden in consanguineous families, contributing to as much as 5-10% of cases. An additional line of evidence pointing to inherited factors for ASD in consanguineous families is the M/F ratio. All data sets show a progressively smaller M/F ratio with additional evidence of familiarity. The most dramatic change in the M/F ratio occurs for families with a single affected child. In this case requiring that any relative, including parents, present with any form of intellectual disability reduce the M/F ratio for consanguineous as well as non-consanguineous families.  However the rate of change is very different.  For consanguineous families M/F drops from 8 to 3 while for non-consanguineous families from 7 to 5.  It is far more likely that close relatives, most frequently parents, cousins and uncles, share just a single allele rather than two causative alleles required for recessive inheritance.  Thus the far smaller impact of additional familiarity requirements in consanguineous families underscores the importance of recessive inheritance in families with high-observed homozygosity.