Genetic influence of allelic variation on the phenotypic variability of cognitive outcomes in individuals with Williams syndrome.

Williams syndrome (WS) is caused by a rare microdeletion of 25 – 27 genes on chromosome 7q11.23, characterized by a distinctive cognitive profile with considerable phenotypic heterogeneity. However, the genetic factors underlying this phenotypic variability remain unclear. I explored two genetic mechanisms—parent-of-origin (PoO) effects and genome-wide polygenic variation, as measured by polygenic scores (PGSs)—to elucidate possible influences on cognitive variability in WS. The first study examined PoO effects on cognitive abilities in two overlapping samples of children with WS (N = 211, N = 251). PoO refers to which parental homolog of chromosome 7 contains the WS deletion. Despite inconsistent findings, some researchers have argued that differential expression of certain genes within the deleted region drive PoO-dependent differences in cognitive outcomes. Using the Differential Ability Scales-II (DAS-II) and the Kaufman Brief Intelligence Test-2 (KBIT-2), I investigated whether PoO significantly affected overall or domain-specific cognitive abilities. For each DAS-II and KBIT-2 measure considered, statistical analyses confirmed that PoO differences were not statistically significant, and, more importantly, that p-values were high and associated effect sizes were negligible, providing strong evidence against PoO-dependent differences in cognitive performance within this population. The second study applied a polygenic score (PGS) representing additive genetic predisposition for educational attainment (EA) to assess how genetic variation across the genome might impact cognitive abilities in individuals with WS. EA PGSs consistently predict significant variability in intellectual outcomes in large typically developing (TD) samples but no previous studies using this approach to address intellectual variability in individuals with WS have been reported. EA PGS did not account for a significant proportion of the variability in overall intellectual ability, verbal ability, or nonverbal reasoning ability in WS. However, EA PGS was significantly associated with spatial abilities, accounting for 3.5% of the variability in spatial performance after controlling for sex (p = 0.03), despite the limited power due to a sample size (N = 132) that although large for a study of individuals with WS was small for PGS-focused research. This finding suggests that PGSs hold promise for understanding cognitive variability in non-TD populations. The research in this dissertation provides strong evidence that PoO effects do not contribute to cognitive variability in individuals with WS. At the same time, it demonstrates the promise of PGS approaches for accounting for a small but significant proportion of that variability, especially for studies that are adequately powered. Further research using a variety of approaches is needed to provide additional insights toward understanding the genetic basis of the cognitive variability observed among individuals with WS.