The social genome of friends and schoolmates in the National Longitudinal Study of Adolescent to Adult Health

Humans tend to form social relationships with others who resemble them. Whether this sorting of like with like arises from historical patterns of migration, meso-level social structures in modern society, or individual-level selection of similar peers remains unsettled. Recent research has evaluated the possibility that unobserved genotypes may play an important role in the creation of homophilous relationships. We extend this work by using data from 5,500 adolescents from the National Longitudinal Study of Adolescent to Adult Health (Add Health) to examine genetic similarities among pairs of friends. Although there is some evidence that friends have correlated genotypes, both at the whole-genome level as well as at trait-associated loci (via polygenic scores), further analysis suggests that meso-level forces, such as school assignment, are a principal source of genetic similarity between friends. We also observe apparent social-genetic effects in which polygenic scores of an individual's friends and schoolmates predict the individual's own educational attainment. In contrast, an individual's height is unassociated with the height genetics of peers.

Genetic Predisposition to Obesity and Medicare Expenditures

Background

The relationship between obesity and health expenditures is not well understood. We examined the relationship between genetic predisposition to obesity measured by a polygenic risk score for body mass index (BMI) and Medicare expenditures.

Methods

Biennial interview data from the Health and Retirement Survey for a nationally representative sample of older adults enrolled in fee-for-service Medicare were obtained from 1991 through 2010 and linked to Medicare claims for the same period and to Genome-Wide Association Study (GWAS) data. The study included 6,628 Medicare beneficiaries who provided 68,627 complete person-year observations during the study period. Outcomes were total and service-specific Medicare expenditures and indicators for expenditures exceeding the 75th and 90th percentiles. The BMI polygenic risk score was derived from GWAS data. Regression models were used to examine how the BMI polygenic risk score was related to health expenditures adjusting for demographic factors and GWAS-derived ancestry.

Results

Greater genetic predisposition to obesity was associated with higher Medicare expenditures. Specifically, a 1 SD increase in the BMI polygenic risk score was associated with a $805 (p < .001) increase in annual Medicare expenditures per person in 2010 dollars (~15% increase), a $370 (p < .001) increase in inpatient expenses, and a $246 (p < .001) increase in outpatient services. A 1 SD increase in the polygenic risk score was also related to increased likelihood of expenditures exceeding the 75th percentile by 18% (95% CI: 10%-28%) and the 90th percentile by 27% (95% CI: 15%-40%).

Conclusion

Greater genetic predisposition to obesity is associated with higher Medicare expenditures.

Genetic Heterogeneity in Depressive Symptoms Following the Death of a Spouse: Polygenic Score Analysis of the U.S. Health and Retirement Study

OBJECTIVE

Experience of stressful life events is associated with risk of depression. Yet many exposed individuals do not become depressed. A controversial hypothesis is that genetic factors influence vulnerability to depression following stress. This hypothesis is often tested with a "diathesis-stress" model, in which genes confer excess vulnerability. The authors tested an alternative formulation of this model: genes may buffer against depressogenic effects of life stress.

METHOD

The hypothesized genetic buffer was measured using a polygenic score derived from a published genome-wide association study of subjective well-being. The authors tested whether married older adults who had higher polygenic scores were less vulnerable to depressive symptoms following the death of their spouse compared with age-matched peers who had also lost their spouse and who had lower polygenic scores. Data were analyzed from 8,588 non-Hispanic white adults in the Health and Retirement Study (HRS), a population-representative longitudinal study of older adults in the United States.

RESULTS

HRS adults with higher well-being polygenic scores experienced fewer depressive symptoms during follow-up. Those who survived the death of their spouses (N=1,647) experienced a sharp increase in depressive symptoms following the death and returned toward baseline over the following 2 years. Having a higher well-being polygenic score buffered against increased depressive symptoms following a spouse's death.

CONCLUSIONS

The effects were small, and the clinical relevance is uncertain, although polygenic score analyses may provide clues to behavioral pathways that can serve as therapeutic targets. Future studies of gene-environment interplay in depression may benefit from focus on genetics discovered for putative protective factors.

Mortality selection in a genetic sample and implications for association studies

Background

Mortality selection occurs when a non-random subset of a population of interest has died before data collection and is unobserved in the data. Mortality selection is of general concern in the social and health sciences, but has received little attention in genetic epidemiology. We tested the hypothesis that mortality selection may bias genetic association estimates, using data from the US-based Health and Retirement Study (HRS).

Methods

We tested mortality selection into the HRS genetic database by comparing HRS respondents who survive until genetic data collection in 2006 with those who do not. We next modelled mortality selection on demographic, health and social characteristics to calculate mortality selection probability weights. We analysed polygenic score associations with several traits before and after applying inverse-probability weighting to account for mortality selection. We tested simple associations and time-varying genetic associations (i.e. gene-by-cohort interactions).

Results

We observed mortality selection into the HRS genetic database on demographic, health and social characteristics. Correction for mortality selection using inverse probability weighting methods did not change simple association estimates. However, using these methods did change estimates of gene-by-cohort interaction effects. Correction for mortality selection changed gene-by-cohort interaction estimates in the opposite direction from increased mortality selection based on analysis of HRS respondents surviving through 2012.

Conclusions

Mortality selection may bias estimates of gene-by-cohort interaction effects. Analyses of HRS data can adjust for mortality selection associated with observables by including probability weights. Mortality selection is a potential confounder of genetic association studies, but the magnitude of confounding varies by trait.

Changing Polygenic Penetrance on Phenotypes in the 20(th) Century Among Adults in the US Population

This study evaluates changes in genetic penetrance—defined as the association between an additive polygenic score and its associated phenotype—across birth cohorts. Situating our analysis within recent historical trends in the U.S., we show that, while height and BMI show increasing genotypic penetrance over the course of 20^th Century, education and heart disease show declining genotypic effects. Meanwhile, we find genotypic penetrance to be historically stable with respect to depression. Our findings help inform our understanding of how the genetic and environmental landscape of American society has changed over the past century, and have implications for research which models gene-environment (GxE) interactions, as well as polygenic score calculations in consortia studies that include multiple birth cohorts.

Genome-Wide Estimates of Heritability for Social Demographic Outcomes

An increasing number of studies that are widely used in the demographic research community have collected genome-wide data from their respondents. It is therefore important that demographers have a proper understanding of some of the methodological tools needed to analyze such data. This article details the underlying methodology behind one of the most common techniques for analyzing genome-wide data, genome-wide complex trait analysis (GCTA). GCTA models provide heritability estimates for health, health behaviors, or indicators of attainment using data from unrelated persons. Our goal was to describe this model, highlight the utility of the model for biodemographic research, and demonstrate the performance of this approach under modifications to the underlying assumptions. The first set of modifications involved changing the nature of the genetic data used to compute genetic similarities between individuals (the genetic relationship matrix). We then explored the sensitivity of the model to heteroscedastic errors. In general, GCTA estimates are found to be robust to the modifications proposed here, but we also highlight potential limitations of GCTA estimates.

The Geographic Distribution of Genetic Risk as Compared to Social Risk for Chronic Diseases in the United States

There is an association between chronic disease and geography, and there is evidence that the environment plays a critical role in this relationship. Yet at the same time, there is known to be substantial geographic variation by ancestry across the United States. Resulting geographic genetic variation-that is, the extent to which single nucleotide polymorphisms (SNPs) related to chronic disease vary spatially-could thus drive some part of the association between geography and disease. We describe the variation in chronic disease genetic risk by state of birth by taking risk SNPs from genome-wide association study meta-analyses for coronary artery disease, diabetes, and ischemic stroke and creating polygenic risk scores. We compare the amount of variability across state of birth in these polygenic scores to the variability in parental education, own education, earnings, and wealth. Our primary finding is that the polygenic risk scores are only weakly differentially distributed across U.S. states. The magnitude of the differences in geographic distribution is very small in comparison to the distribution of social and economic factors and thus is not likely sufficient to have a meaningful effect on geographic disease differences by U.S. state.

Breastfeeding is associated with waist-to-height ratio in young adults

Background

The current study investigated the association between breastfeeding and adult weight distribution using an emerging indicator of weight distribution, the waist-to-height ratio (WHtR).

Methods

The study sample consisted of two subsamples of individuals that were part of the National Longitudinal Study of Adolescent Health. One sample (n = 1 179) consisted of individuals from the sibling pair data. A second sample (n = 4 648) consisted of individuals that were not part of the paired data. Regression models were constructed to establish if there was a relationship between breastfeeding and two measures of weight distribution: WHtR and body mass index (BMI). Controls for parental socioeconomic status, maternal smoking, race, sex, age, birth weight, maternal BMI, genetic ancestry, and a genetic risk score (GRS) for obesity were included. In addition, a behavioral risk score (BRS) was constructed to control for other residual confounding factors.

Results

A significant, inverse relationship between breastfeeding and adult WHtR persisted in models constructed from the sibling pair sample (P = 0.002) and unrelated sample (P <0.0001). This association remained significant with the inclusion of ancestry principal components, GRS, and a measure of maternal obesity.

Conclusions

The moderate association between breastfeeding and weight distribution persists into adulthood while controlling for potential confounders. This paper also provides evidence that the WHtR may be a superior outcome measure to BMI in studies investigating breastfeeding and obesity.

Electronic supplementary material

The online version of this article (doi:10.1186/s12889-015-2611-7) contains supplementary material, which is available to authorized users.

Cohort Effects in the Genetic Influence on Smoking

We examine the hypothesis that the heritability of smoking has varied over the course of recent history as a function of associated changes in the composition of the smoking and non-smoking populations. Classical twin-based heritability analysis has suggested that genetic basis of smoking has increased as the information about the harms of tobacco has become more prevalent-particularly after the issuance of the 1964 Surgeon General's Report. In the present paper we deploy alternative methods to test this claim. We use data from the Health and Retirement Study to estimate cohort differences in the genetic influence on smoking using both genomic-relatedness-matrix restricted maximum likelihood and a modified DeFries-Fulker approach. We perform a similar exercise deploying a polygenic score for smoking using results generated by the Tobacco and Genetics consortium. The results support earlier claims that the genetic influence in smoking behavior has increased over time. Emphasizing historical periods and birth cohorts as environmental factors has benefits over existing GxE research. Our results provide additional support for the idea that anti-smoking policies of the 1980s may not be as effective because of the increasingly important role of genotype as a determinant of smoking status.

Polygenic Influence on Educational Attainment: New evidence from The National Longitudinal Study of Adolescent to Adult Health

Recent studies have begun to uncover the genetic architecture of educational attainment. We build on this work using genome-wide data from siblings in the National Longitudinal Study of Adolescent to Adult Health (Add Health). We measure the genetic predisposition of siblings to educational attainment using polygenic scores. We then test how polygenic scores are related to social environments and educational outcomes. In Add Health, genetic predisposition to educational attainment is patterned across the social environment. Participants with higher polygenic scores were more likely to grow up in socially advantaged families. Even so, the previously published genetic associations appear to be causal. Among pairs of siblings, the sibling with the higher polygenic score typically went on to complete more years of schooling as compared to their lower-scored co-sibling. We found subtle differences between sibling fixed effect estimates of the genetic effect versus those based on unrelated individuals.

Prevention, Use of Health Services, and Genes: Implications of Genetics for Policy Formation

We evaluate the hypothesis that genetic factors influence the use of health services and prevention behaviors in a national sample of adult twins in the United States. The analysis compares the correlation of these outcomes between identical twins, who share all their genes, to the correlation between nonidentical twins, who share, on average, only one-half of their genes. Because the environmental similarities of twins are assumed to be the same for identical and nonidentical twin pairs, researchers can partition the variance in behavioral outcomes that are due to genetic and environmental factors. Using established methods in this field, we find evidence of significant genetic influences on preferences toward prevention, overall prevention effort, routine checkups, and prescription drug use. Use of curative services does not appear to be influenced by genes. Our findings offer several implications for policymakers and researchers and suggest that genetics could be informative for health services and policy research.

Is the Effect of Parental Education on Offspring Biased or Moderated by Genotype?

Parental education is the strongest measured predictor of offspring education, and thus many scholars see the parent-child correlation in educational attainment as an important measure of social mobility. But if social changes or policy interventions are going to have dynastic effects, we need to know what accounts for this intergenerational association, that is, whether it is primarily environmental or genetic in origin. Thus, to understand whether the estimated social influence of parental education on offspring education is biased owing to genetic inheritance (or moderated by it), we exploit the findings from a recent large genome-wide association study of educational attainment to construct a genetic score designed to predict educational attainment. Using data from two independent samples, we find that our genetic score significantly predicts years of schooling in both between-family and within-family analyses. We report three findings that should be of interest to scholars in the stratification and education fields. First, raw parent-child correlations in education may reflect one-sixth genetic transmission and five-sixths social inheritance. Second, conditional on a child's genetic score, a parental genetic score has no statistically significant relationship to the child's educational attainment. Third, the effects of offspring genotype do not seem to be moderated by measured sociodemographic variables at the parental level (but parent-child genetic interaction effects are significant). These results are consistent with the existence of two separate systems of ascription: genetic inheritance (a random lottery within families) and social inheritance (across-family ascription). We caution, however, that at the presently attainable levels of explanatory power, these results are preliminary and may change when better-powered genetic risk scores are developed.

The National Longitudinal Study of Adolescent to Adult Health (Add Health) sibling pairs genome-wide data

Here we provide a detailed description of the genome-wide information available on the National Longitudinal Study of Adolescent to Adult Health (Add Health) sibling pair subsample (Harris et al. in Twin Res Hum Genet 16:391-398, 2013). A total of 2,020 samples were genotyped (including duplicates) arising from 1946 Add Health individuals from the sibling pairs subsample. After various steps for quality control (QC) and quality assurance (QA), we have high quality genome-wide data available on 1,888 individuals. In this report, we first highlight the QC and QA steps that were taken to prune the data of poorly performing samples and genetic markers. We further estimate the pairwise biological relationships using genome-wide data and compare those estimates to the assumed relationships in Add Health. Additionally, using genome-wide data from known regional reference populations from Europe, West Africa, North and South America, Japan and China, we estimate the relative genetic ancestry of the respondents. Finally, rather than conducting a traditional cross-sectional genome-wide association study (GWAS) of body mass index (BMI), we opted to utilize the extensive publicly available genome-wide information to conduct a weighted GWAS of longitudinal BMI while accounting for both family and ethnic variation.

Replicability and robustness of genome-wide-association studies for behavioral traits

A recent genome-wide-association study of educational attainment identified three single-nucleotide polymorphisms (SNPs) whose associations, despite their small effect sizes (each R (2) ≈ 0.02%), reached genome-wide significance (p < 5 × 10(-8)) in a large discovery sample and were replicated in an independent sample (p < .05). The study also reported associations between educational attainment and indices of SNPs called "polygenic scores." In three studies, we evaluated the robustness of these findings. Study 1 showed that the associations with all three SNPs were replicated in another large (N = 34,428) independent sample. We also found that the scores remained predictive (R (2) ≈ 2%) in regressions with stringent controls for stratification (Study 2) and in new within-family analyses (Study 3). Our results show that large and therefore well-powered genome-wide-association studies can identify replicable genetic associations with behavioral traits. The small effect sizes of individual SNPs are likely to be a major contributing factor explaining the striking contrast between our results and the disappointing replication record of most candidate-gene studies.

Norms as Group-Level Constructs: Investigating School-Level Teen Pregnancy Norms and Behaviors

Social norms are a group-level phenomenon, but past quantitative research has rarely measured them in the aggregate or considered their group-level properties. We used the school-based design of the National Longitudinal Study of Adolescent Health to measure normative climates regarding teen pregnancy across 75 U.S. high schools. We distinguished between the strength of a school's norm against teen pregnancy and the consensus around that norm. School-level norm strength and dissensus were strongly (r = -0.65) and moderately (r = 0.34) associated with pregnancy prevalence within schools, respectively. Normative climate partially accounted for observed racial differences in school pregnancy prevalence, but norms were a stronger predictor than racial composition. As hypothesized, schools with both a stronger average norm against teen pregnancy and greater consensus around the norm had the lowest pregnancy prevalence. Results highlight the importance of group-level normative processes and of considering the local school environment when designing policies to reduce teen pregnancy.

Polygenic risk predicts obesity in both white and black young adults

OBJECTIVE

To test transethnic replication of a genetic risk score for obesity in white and black young adults using a national sample with longitudinal data.

DESIGN AND METHODS

A prospective longitudinal study using the National Longitudinal Study of Adolescent Health Sibling Pairs (n = 1,303). Obesity phenotypes were measured from anthropometric assessments when study members were aged 18-26 and again when they were 24-32. Genetic risk scores were computed based on published genome-wide association study discoveries for obesity. Analyses tested genetic associations with body-mass index (BMI), waist-height ratio, obesity, and change in BMI over time.

RESULTS

White and black young adults with higher genetic risk scores had higher BMI and waist-height ratio and were more likely to be obese compared to lower genetic risk age-peers. Sibling analyses revealed that the genetic risk score was predictive of BMI net of risk factors shared by siblings. In white young adults only, higher genetic risk predicted increased risk of becoming obese during the study period. In black young adults, genetic risk scores constructed using loci identified in European and African American samples had similar predictive power.

CONCLUSION

Cumulative information across the human genome can be used to characterize individual level risk for obesity. Measured genetic risk accounts for only a small amount of total variation in BMI among white and black young adults. Future research is needed to identify modifiable environmental exposures that amplify or mitigate genetic risk for elevated BMI.

Understanding multiple levels of norms about teen pregnancy and their relationships to teens' sexual behaviors

Researchers seeking to understand teen sexual behaviors often turn to age norms, but they are difficult to measure quantitatively. Previous work has usually inferred norms from behavioral patterns or measured group-level norms at the individual level, ignoring multiple reference groups. Capitalizing on the multilevel design of the Add Health survey, we measure teen pregnancy norms perceived by teenagers, as well as average norms at the school and peer network levels. School norms predict boys' perceived norms, while peer network norms predict girls' perceived norms. Peer network and individually perceived norms against teen pregnancy independently and negatively predict teens' likelihood of sexual intercourse. Perceived norms against pregnancy predict increased likelihood of contraception among sexually experienced girls, but sexually experienced boys' contraceptive behavior is more complicated: When both the boy and his peers or school have stronger norms against teen pregnancy he is more likely to contracept, and in the absence of school or peer norms against pregnancy, boys who are embarrassed are less likely to contracept. We conclude that: (1) patterns of behavior cannot adequately operationalize teen pregnancy norms, (2) norms are not simply linked to behaviors through individual perceptions, and (3) norms at different levels can operate independently of each other, interactively, or in opposition. This evidence creates space for conceptualizations of agency, conflict, and change that can lead to progress in understanding age norms and sexual behaviors.

Testing the key assumption of heritability estimates based on genome-wide genetic relatedness

Comparing genetic and phenotypic similarity among unrelated individuals seems a promising way to quantify the genetic component of traits while avoiding the problematic assumptions plaguing twin- and other kin-based estimates of heritability. One approach uses a Genetic Relatedness Estimation through Maximum Likelihood (GREML) model for individuals who are related at less than 0.025 to predict their phenotypic similarity by their genetic similarity. Here we test the key underlying assumption of this approach: that genetic relatedness is orthogonal to environmental similarity. Using data from the Health and Retirement Study (and two other surveys), we show two unrelated individuals may be more likely to have been reared in a similar environment (urban versus nonurban setting) if they are genetically similar. This effect is not eliminated by controls for population structure. However, when we include this environmental confound in GREML models, heritabilities do not change substantially and thus potential bias in estimates of most biological phenotypes is probably minimal.

Gene-environment interactions related to body mass: School policies and social context as environmental moderators

This paper highlights the role of institutional resources and policies, whose origins lie in political processes, in shaping the genetic etiology of body mass among a national sample of adolescents. Using data from Waves I and II of the National Longitudinal Study of Adolescent Health, we decompose the variance of body mass into environmental and genetic components. We then examine the extent to which the genetic influences on body mass are different across the 134 schools in the study. Taking advantage of school differences in both health-related policies and social norms regarding body size, we examine how institutional resources and policies alter the relative impact of genetic influences on body mass. For the entire sample, we estimate a heritability of .82, with the remaining .18 due to unique environmental factors. However, we also show variation about this estimate and provide evidence suggesting that social norms and institutional policies often mask genetic vulnerabilities to increased weight. Empirically, we demonstrate that more-restrictive school policies and policies designed to curb weight gain are also associated with decreases the proportion of variance in body mass that is due to additive genetic influences.

Ethnicity, body mass, and genome-wide data

This article combines social and genetic epidemiology to examine the influence of self-reported ethnicity on body mass index (BMI) among a sample of adolescents and young adults. We use genetic information from more than 5,000 single nucleotide polymorphisms in combination with principal components analysis to characterize population ancestry of individuals in this study. We show that non-Hispanic white and Mexican-American respondents differ significantly with respect to BMI and differ on the first principal component from the genetic data. This first component is positively associated with BMI and accounts for roughly 3% of the genetic variance in our sample. However, after controlling for this genetic measure, the observed ethnic differences in BMI remain large and statistically significant. This study demonstrates a parsimonious method to adjust for genetic differences among individual respondents that may contribute to observed differences in outcomes. In this case, adjusting for genetic background has no bearing on the influence of self-identified ethnicity.