Understanding the genetic basis of phenotype variability in individuals with neurocognitive disorders

Thesis (Ph.D.)--University of Washington, 2016-06

Individuals with a diagnosis of a neurocognitive disorder, such as an autism spectrum disorder (ASD), can present with a wide range of phenotypes. Some have severe language and cognitive deficiencies while others are only deficient in social functioning. Sequencing studies have revealed extreme locus heterogeneity underlying the ASDs. Even cases with a known pathogenic variant, such as the 16p11.2 CNV, can be associated with phenotypic heterogeneity. In this thesis, I test the hypothesis that phenotypic heterogeneity observed in populations with a known pathogenic variant, such as the 16p11.2 CNV as well as that associated with the ASDs in general, is due to additional genetic factors. I analyze the phenotypic and genotypic characteristics of over 120 families where at least one individual carries the 16p11.2 CNV, as well as a cohort of over 40 families with high functioning autism and/or intellectual disability. In the 16p11.2 cohort, I assessed variation both internal to and external to the CNV critical region. Among de novo cases, I found a strong maternal bias for the origin of deletions (59/66, 89.4% of cases, p=2.38x10^-11), the strongest such effect so far observed for a CNV associated with a microdeletion syndrome, a significant maternal transmission bias for secondary deletions (32 maternal versus 14 paternal, p=1.14x10^-2), and nine probands carrying additional CNVs disrupting autism-associated genes. In the same cohort, I assessed genome wide exonic variation, including in the 27 16p11.2 CNV critical region genes and the 3 genes that lie in the flanking segmental duplications, BOLA2, SLX1A, and SULT1A3 with the hypothesis that dosage imbalance in these genes could lead to variable phenotypes. I find an absence of variation across the critical region, compared to similarly sized regions genome-wide by average heterozygosity (2nd percentile) and Tajima’s D (3rd percentile) metrics. Among the 27 critical region genes and three duplicated genes, I find no loss of function variants in 16p11.2 CNV carriers. Our genome-wide exome analysis revealed 13 likely-gene disruptive (LGD) variants in 13 probands in autism-associated genes, which is fewer than would be expected by chance (p<10^-16) and individuals having such variants trend towards being more severely affected on FSIQ (p=0.19). To understand the genetic heterogeneity associated with high-functioning autism and intellectual disability, I assessed genetic variation observed in a cohort of 43 local families of which 29 have a diagnosis of high functioning-autism. I discovered variants in novel autism candidate genes, including LPHN1 and NUMBL, find that the high functioning autism cohort tends to have more inherited loss of function and severe missense variation per individual than low functioning cohorts (p<2.2x10^-16), but fewer de novo LGD variants per individual (p=0.007). I also find that de novo variants in high functioning cases lie in a protein-protein interaction network including proteins involved in the NOTCH signaling pathway. Our findings suggest that modifiers external to, as opposed to variants internal to the critical region, may play a role in the observed phenotypic differences observed in individuals with a 16p11.2 CNV and those with ASDs in general.