The Search is On for Missing Heritability -- and Missing "Environmentality" -- of Personality: A Developmental Approach

Abstract

Twin and adoption research has consistently found evidence that genetic and unique environmental factors each account for about half of the observed individual differences in personality traits. However, studies aimed at finding associated genetic polymorphisms have had limited success. The inability to link genetic variants to personality phenotypes has become commonly known as the “missing heritability” problem (Manolio et al., 2009). Similarly, Bleidorn, Kandler, and Caspi (2014) recently named the difficulty in identifying specific environmental factors as causal influences on personality phenotypes the problem of “missing environment.” One factor that may contribute to the difficulty in uncovering specific influences on personality is that the phenotype changes over time. In study 1, I performed bivariate Cholesky decomposition of personality traits across ages 14 and 29; genetic influences on personality were largely stable over this developmental period, but not entirely so (mean rg = .75). Next, I examined the extent to which our assumptions about the additive nature of genetic influence on personality are true by performing univariate Genome-Wide Complex Trait Analysis (GCTA) separately at ages 17, 24, and 29. Restricting GCTA to specific developmental groups did not produce estimates of SNP heritability that were consistently different from those based on an adult sample. Significant SNP heritabilities were observed for the MPQ primary scales aggression, harm avoidance, and traditionalism as well as for the super-factor constraint; results suggested that SNPs could account for about half of the twin-estimated heritability for these traits (~20%). In study 2, I investigated the effects of getting married and having children on personality development across ages 17 and 29; after demonstrating a small but significant relationship between these role transitions and MPQ super-factors constraint and negative emotionality, I followed up with a co-twin control analysis to determine whether such effects were causal. There were no significant differences in personality change between identical twins discordant for these role transitions, suggesting that associations could be attributed to familial factors underlying both personality and initiation of adult social roles.