Nature via Nurture: Using novel methods to shed new light on how gene-environment interactions shape social, economic and political attitudes and behavior
Why are some people more likely to be highly educated, end up in high status jobs, be healthier and live longer, be more politically engaged, and, in general, happier than others? Understanding the origins of such social, economic and political inequalities in life course outcomes between individuals is of fundamental importance for the social sciences. Recent interdisciplinary research suggests that complex human behaviors and life chances may be shaped by an intricate interplay between genes and environment. However, the knowledge of how these so called gene-by-environment interactions actually work is limited: what genetic factors interact with what social, economic and political factors, and how?
The dearth of stable and reproducible progress in this area has in large parts been due to two severe problems: limitations in the available scientific methods and shortage of adequate data. In this project we intend to overcome these problems by making use of recent innovations in molecular genetics and methodological advances in the social sciences in combination with unique Swedish register data. Thus, the goal of the project is to investigate in a robust and credible way how gene-environment interactions shape social, economic, and political attitudes and behavior. In the end, a more solid foundation for the research on gene-environment interactions can also aid in developing more effective policies that deal with the social roots and consequences of inequality.
Final report
The purpose of the project is to robustly and credibly investigate how social, economic, and political attitudes and behaviors are shaped by the interplay between genetic and social factors. Why do some people attain higher education and better jobs, lead healthier and longer lives, are more politically engaged, and generally happier than others? A better understanding of the causes behind such social, economic, and political inequalities is central to several social science disciplines. Contemporary research suggests that similar complex human behaviors and life chances may be shaped by an intricate interaction between genes and the environment. However, the knowledge of how these so-called gene-environment interactions actually manifest is very limited: which genetic factors interact with which social, economic, and political factors, and how?
The lack of stable and reproducible progress in this field has largely been due to two formidable obstacles: limitations in the available scientific methods and the lack of suitable data. The aim of this project has been to address these problems by leveraging several innovations from molecular genetics and methodological advances in the social sciences, combined with unique registry data from Sweden and other countries. A better and deeper understanding of these issues can also improve the ability to develop more effective policy measures to counteract societal inequality.
The project has progressed in accordance with the plan during the project period. Much of the work in the first year focused on compiling the data needed to address the project's research questions. We have built a unique dataset centered around all individuals in the Swedish Twin Registry (STR) who also have genetic information in the form of millions of genetic markers. These individuals have been linked to their parents, grandparents, partners, (non-genotyped) siblings, and children. This entire population has been connected to relevant registry information from Statistics Sweden (e.g., education level, income, and employment status), the National Board of Health and Welfare (various health outcomes), the Swedish National Council for Crime Prevention (information on convictions), the Swedish Armed Forces (conscription tests), and information from the Twin Registry's own surveys (e.g., health outcomes, personality variables, political and economic attitudes and behaviors). Additionally, we have secured access to several similar datasets in the USA and England.
The core of our datasets is the so-called polygenic indices (PGIs) that we construct based on large genetic association studies (GWAS) and then link to individuals in various study cohorts (e.g., individuals in the STR). A PGI is designed to capture as much of the genetic variation in a specific outcome as possible. This means that we can create separate PGIs for different outcomes. For example, our PGI for educational attainment can explain nearly 15% of the variation in educational level among individuals. Constructing these polygenic indices is very time-consuming. An important part of the project has been to make this unique data publicly available so that not only we but also anyone interested in using genetic information in the form of polygenic indices in their research can do so. This was accomplished in 2021 when we simultaneously published an article in Nature Human Behavior describing this dataset and how it should be used. Work on these PGIs continues, and later this year, we will release an updated version of the database with PGIs for individuals in a large number of existing biobanks and cohorts of interest for social science research.
Within the framework of the project, we have conducted and contributed to several very large GWAS studies to map the relationship between millions of genetic markers and various life outcomes (e.g., health, personality, and education). Many of these have formed the basis for some of the polygenic indices described above. These studies were based on collaborations within larger consortia with contributions from a large number of research groups worldwide. Several of these studies have been led and coordinated by members of the project team. An example is the article “Polygenic prediction of educational attainment within and between families from genome-wide association analyses in 3 million individuals” published in Nature Genetics, which examines genetic effects on education. This study includes an analysis of more than 3,000,000 individuals and was, at the time of publication, the most extensive GWAS study conducted on any outcome.
An additional primary objective of the project has been to conduct studies of gene-environment interactions—situations where the effect of genetic factors on people's attitudes and behaviors depends on the social context, and vice versa, where the effect of social factors depends on individuals' genetic predispositions. Here too, we have completed several studies. One example is an investigation of how the Swedish compulsory school reform during the 1950s and 1960s affected individuals differently depending on their genetic predispositions. The reform extended mandatory schooling from 7 to 9 years and was implemented in a way that allows for causal estimates of being subjected to a new school system. The working hypothesis was that the reform had equalizing effects in such a way that the positive effects of extended schooling on various life outcomes (e.g., higher education, income, and health) are stronger among individuals with poorer genetic predispositions for the same outcomes. The results show that this is partly the case. In another study, we examine how genetic predispositions for education influence the willingness to vote in general elections. The results here show that genetic factors significantly affect the propensity to vote and that these effects are more pronounced in elections perceived by voters as less important (e.g., European Parliament elections in Sweden or midterm elections in the USA).
During the project, we also initiated a collaboration with researchers at the University of Bristol, aiming to conduct several large studies on genetic effects on health and welfare. These studies are based on so-called within-family models, where genetic effects are estimated by using differences in genotypes and various health and welfare outcomes between individuals in sibling pairs. The advantage of this approach is that the genetic effects we discover can be causally interpreted with high confidence.
Several of the studies we have conducted make significant contributions to the international research frontier. Firstly, the genetic association studies we have conducted and participated in are the largest of their kind, providing more precise estimates of genetic effects on various life outcomes. Secondly, the data we have access to have enabled much more stringent estimates of causal gene-environment interactions. Thirdly, the bank of polygenic indices we have published is an important contribution in itself as it facilitates other researchers to incorporate genetic information into their research.
Regarding the communication of research results, we have primarily targeted reputable and peer-reviewed journals of both general and subject-specific nature. As a rule, we publish our reports on so-called preprint servers (most often bioRxiv) when we submit them for review to an appropriate journal. Members of the project team regularly present results from the various studies in the project at various national and international conferences. Additionally, on several occasions during the project period, we have been invited to present results from our research in other settings. It is also worth mentioning that we publish popular science summaries of our major studies on the following website: https://www.thessgac.org/faqs. These summaries are intended for the interested public and media and include detailed discussions on how to interpret the results we present in the studies, as well as how not to interpret them.
The project work has also been integrated into the Department of Government in Uppsala. In particular, we have had the opportunity to recruit two doctoral students who commenced their studies in 2020 and 2021, respectively. The doctoral positions have not been funded by RJ. However, both doctoral students are writing their dissertations within the framework of the project's overarching objectives.
The lack of stable and reproducible progress in this field has largely been due to two formidable obstacles: limitations in the available scientific methods and the lack of suitable data. The aim of this project has been to address these problems by leveraging several innovations from molecular genetics and methodological advances in the social sciences, combined with unique registry data from Sweden and other countries. A better and deeper understanding of these issues can also improve the ability to develop more effective policy measures to counteract societal inequality.
The project has progressed in accordance with the plan during the project period. Much of the work in the first year focused on compiling the data needed to address the project's research questions. We have built a unique dataset centered around all individuals in the Swedish Twin Registry (STR) who also have genetic information in the form of millions of genetic markers. These individuals have been linked to their parents, grandparents, partners, (non-genotyped) siblings, and children. This entire population has been connected to relevant registry information from Statistics Sweden (e.g., education level, income, and employment status), the National Board of Health and Welfare (various health outcomes), the Swedish National Council for Crime Prevention (information on convictions), the Swedish Armed Forces (conscription tests), and information from the Twin Registry's own surveys (e.g., health outcomes, personality variables, political and economic attitudes and behaviors). Additionally, we have secured access to several similar datasets in the USA and England.
The core of our datasets is the so-called polygenic indices (PGIs) that we construct based on large genetic association studies (GWAS) and then link to individuals in various study cohorts (e.g., individuals in the STR). A PGI is designed to capture as much of the genetic variation in a specific outcome as possible. This means that we can create separate PGIs for different outcomes. For example, our PGI for educational attainment can explain nearly 15% of the variation in educational level among individuals. Constructing these polygenic indices is very time-consuming. An important part of the project has been to make this unique data publicly available so that not only we but also anyone interested in using genetic information in the form of polygenic indices in their research can do so. This was accomplished in 2021 when we simultaneously published an article in Nature Human Behavior describing this dataset and how it should be used. Work on these PGIs continues, and later this year, we will release an updated version of the database with PGIs for individuals in a large number of existing biobanks and cohorts of interest for social science research.
Within the framework of the project, we have conducted and contributed to several very large GWAS studies to map the relationship between millions of genetic markers and various life outcomes (e.g., health, personality, and education). Many of these have formed the basis for some of the polygenic indices described above. These studies were based on collaborations within larger consortia with contributions from a large number of research groups worldwide. Several of these studies have been led and coordinated by members of the project team. An example is the article “Polygenic prediction of educational attainment within and between families from genome-wide association analyses in 3 million individuals” published in Nature Genetics, which examines genetic effects on education. This study includes an analysis of more than 3,000,000 individuals and was, at the time of publication, the most extensive GWAS study conducted on any outcome.
An additional primary objective of the project has been to conduct studies of gene-environment interactions—situations where the effect of genetic factors on people's attitudes and behaviors depends on the social context, and vice versa, where the effect of social factors depends on individuals' genetic predispositions. Here too, we have completed several studies. One example is an investigation of how the Swedish compulsory school reform during the 1950s and 1960s affected individuals differently depending on their genetic predispositions. The reform extended mandatory schooling from 7 to 9 years and was implemented in a way that allows for causal estimates of being subjected to a new school system. The working hypothesis was that the reform had equalizing effects in such a way that the positive effects of extended schooling on various life outcomes (e.g., higher education, income, and health) are stronger among individuals with poorer genetic predispositions for the same outcomes. The results show that this is partly the case. In another study, we examine how genetic predispositions for education influence the willingness to vote in general elections. The results here show that genetic factors significantly affect the propensity to vote and that these effects are more pronounced in elections perceived by voters as less important (e.g., European Parliament elections in Sweden or midterm elections in the USA).
During the project, we also initiated a collaboration with researchers at the University of Bristol, aiming to conduct several large studies on genetic effects on health and welfare. These studies are based on so-called within-family models, where genetic effects are estimated by using differences in genotypes and various health and welfare outcomes between individuals in sibling pairs. The advantage of this approach is that the genetic effects we discover can be causally interpreted with high confidence.
Several of the studies we have conducted make significant contributions to the international research frontier. Firstly, the genetic association studies we have conducted and participated in are the largest of their kind, providing more precise estimates of genetic effects on various life outcomes. Secondly, the data we have access to have enabled much more stringent estimates of causal gene-environment interactions. Thirdly, the bank of polygenic indices we have published is an important contribution in itself as it facilitates other researchers to incorporate genetic information into their research.
Regarding the communication of research results, we have primarily targeted reputable and peer-reviewed journals of both general and subject-specific nature. As a rule, we publish our reports on so-called preprint servers (most often bioRxiv) when we submit them for review to an appropriate journal. Members of the project team regularly present results from the various studies in the project at various national and international conferences. Additionally, on several occasions during the project period, we have been invited to present results from our research in other settings. It is also worth mentioning that we publish popular science summaries of our major studies on the following website: https://www.thessgac.org/faqs. These summaries are intended for the interested public and media and include detailed discussions on how to interpret the results we present in the studies, as well as how not to interpret them.
The project work has also been integrated into the Department of Government in Uppsala. In particular, we have had the opportunity to recruit two doctoral students who commenced their studies in 2020 and 2021, respectively. The doctoral positions have not been funded by RJ. However, both doctoral students are writing their dissertations within the framework of the project's overarching objectives.