Getting 'Under the Skin': Human Social Genomics in the Multi-Ethnic Study of Atherosclerosis

摘要

The field of human social genomics examines the role of gene expression as a biological mediator in the relationship between the social environment and health. As many studies in this field have been small, lacked replication, and a consensus on best methodological approach, a larger epidemiologic investigation of such associations was warranted. The overarching goal of this dissertation is to contribute to the human social genomics literature by investigating the association between a range of social environmental factors (i.e. adult and childhood socioeconomic status (SES), loneliness, major or lifetime discrimination, perceived stress, chronic burden, and social support) and monocyte gene expression using the Multi-Ethnic Study of Atherosclerosis (MESA) and three different methodological approaches (i.e. an omnibus test for the entire gene expression set using Global Analysis of Covariance, testing each pair of social environmental factors and gene expression using multiple linear regression, and a machine learning method that performs variable selection with a correlated set of predictors called elastic net).;Aim 1 provides evidence that some genes are "socially sensitive" (i.e. demonstrate differential expression across social environmental exposures). Of the 1,854 gene transcripts previously identified as part of the 'conserved transcriptional response to adversity' only a small percent---between 0 to 11%---were associated with one of the seven aforementioned exposures. In Aim 2, a focus specifically on expression of inflammation and immune response gene pathways based on Gene Ontology classifications found significant associations between loneliness (p=0.003), chronic burden (p=0.002), and major or lifetime discrimination (p=0.045) in global analyses with expression of the small subset of 20 gene transcripts related to the chronic inflammation pathway. In Aim 3, the extent to which social environmental factors explain the relationship between race/ethnicity and gene expression was examined. Of the 1407 unique gene transcripts examined, there were Black/White differences in 32%, Black/Hispanic differences in 19%, and Hispanic/White differences in 24%. Accounting for a racially/ethnically pattered social environmental factor (i.e. discrimination, adult SES) in the association between race/ethnicity and gene expression changed the effect estimate by >10% for less than 5% of the inflammation and immune response genes investigated.;In investigating the associations between social environmental factors and monocyte gene expression, this dissertation provides an important contribution in the field of human social genomics by examining the reproducibility of associations, the utility of different methodological approaches, and its contribution to racial/ethnic differences in gene expression.