INSIGHTS Downloaded from http://science.sciencemag.org/

on February 1, 2018

PERSPECTIVES

GENETICS Using nature to understand nurture Genetic associations show how parenting matters for children’s education

By Philipp D. Koellinger1 that parents provide for their children)? ited family environments influence chil- and K. Paige Harden2 Answering this enduring question can di- dren’s educational success, a phenomenon rectly inform our efforts to reduce social termed genetic nurture. hildren resemble their parents in inequality and disease burden. On page Specifically, Kong et al. show that the health, wealth, and well-being. Is 424 of this issue, Kong et al. (1) use genetic part of the parental genotype that children parent-child similarity in traits and data from trios of parents and offspring behaviors due to nature (the genes to address this question in an intriguing 1De Boelelaan 1085, 1081 HV Amsterdam, Netherlands. that children inherit from their par- way. By measuring parents’ and children’s 2108 E. Dean Keeton Stop #A8000, Austin, TX 78704, USA. C ents) or nurture (the environment genes, they provide evidence that inher- Email: [email protected]; [email protected] IMAGES IMAGES/GETTY HERO PHOTO: 386 26 JANUARY 2018 • VOL 359 ISSUE 6374 sciencemag.org SCIENCE

Published by AAAS

DA_0126Perspectives.indd 386 1/24/18 10:59 AM Genetic nurture considers the influence other compelling example of how tightly Yet, genetic nurture does not undermine of family environment on offspring. genetic and environmental mechanisms the value of genetic associations for pre- are entangled. diction purposes. For example, researchers do not inherit can nonetheless predict Genetic nurture also raises important who want to control for genetic effects in children’s educational attainment. This methodological issues for future stud- order to obtain more precise estimates of genetic nurture effect is an indirect link ies. Geneticists have long been concerned an environmental or medical treatment in between parental genotypes and children’s with population stratification (9), which a randomized controlled trial (12) do not characteristics, not caused by the chil- is defined as the existence of systematic care why the genes are correlated with the dren’s own biology but rather by the family differences in gene frequency between outcome. All that matters is that the genes environment that covaries with parental groups that might also differ in environ- used as control variables predict as much genes. The concept of genetic nurture in ment and culture. Population stratification variance as possible in the target outcome. human families extends previous work by can induce a genotype-phenotype correla- Furthermore, for many scientists nur- behavioral ecologists and animal breeders tion that is, in fact, due to an unidentified ture, not nature, is the phenomenon of in- on indirect genetic effects: environmen- group environment. The classic example terest. The study from Kong et al. provides tally mediated effects of one organism’s is the chopsticks gene (10): Any genetic these scientists with a fascinating new tool genes on the phenotype of a related organ- variant that differs in frequency between for investigating the effects of the family ism (2). For instance, one source of genetic Asians and Europeans would come to be environment. Until now, the primary tools variation in dung beetle body size is how correlated with using chopsticks if popu- to disentangle the effects of a parent’s genes large a brood mass the mother beetle pro- lation stratification were uncontrolled. from their actual parenting were adoption

duces (3). Interestingly, given how much of Controlling for population stratification is studies (13) and children-of-twins studies Downloaded from the animal research on indirect genetic ef- typically accomplished by using statistical (14). However, collecting such data is dif- fects has centered on offspring body ficult, and these study designs rarely size, Kong et al. found no evidence of represent the entire range of environ- genetic nurture on children’s height ments. Kong et al. capitalize on the and body mass index (BMI). For hu- “...genetic nurture provides another same logic as an adoption study. The mans living in wealthy societies, ge- nontransmitted parental alleles func- compelling example of how http://science.sciencemag.org/ netic nurture might be more relevant tion like the genotype of an adoptive for behavior and social achievements tightly genetic and environmental parent, in that they help to shape the than for more biologically proximal mechanisms are entangled.” rearing environment but are indepen- outcomes such as body size. dent from the offspring’s genotype. Although the analyses of Kong et al. Data sets with genotyped trios and focus on a single child within a nuclear fam- techniques that adjust for ancestry-specific high-quality measurements of the family ily, they are careful to remind us that family genetic differences among people. How- environment are, unfortunately, still rare. is more than one generation deep and often ever, there is no clear line between popu- But because of low genotyping costs, the includes more than one child per generation. lation stratification and genetic nurture, trio design developed by Kong et al. could The environment that parents provide for particularly once we realize that genetic become a valuable and cost-effective re-

their children could reflect the long arm of nurture effects might themselves repro- search paradigm for scientists interested on February 1, 2018 nurture by previous ancestors. And, siblings duce across generations: At what point in understanding the impact of family en- might create environments for each other. does family history become ancestry? vironments on human flourishing. Genetic nurture could operate through any For researchers whose aim is to identify The ingenious analysis of genetic trio physical or social environment woven by ge- only those genetic effects that could be data by Kong et al. reminds us yet again of netic kin—a tangled web indeed. causally manipulated by changing an indi- the methodological problems that plague The phenomenon of genetic nurture vidual’s own genotype, genetic nurture is, scientists as we try to understand individ- underscores that the results of genetic as- like population stratification, a source of ual differences in complex human behav- sociation studies cannot be interpreted as false results. iors and diseases—but also illustrates how support for a biologically determinist ac- Genetic nurture also presents a new chal- understanding nature can provide us with count of human individual differences, lenge for research designs such as Mende- new tools for studying nurture. j as most social scientists who work with lian randomization (11) that aim to identify genetic data have stressed (4, 5). Rather, causal effects by using genes as if they were REFERENCES 1. A. Kong et al., Science 359, 424 (2018). results from genetic association studies naturally occurring experiments. For ex- 2. J. B. Wolf et al., Trends Ecol. Evol. 13, 64 (1998). are correlations that might point the way ample, a researcher might use the alleles 3. J. Hunt, L. W. Simmons, Proc. Natl. Acad. Sci. U.S.A. 99, toward understanding causal mechanisms, associated with educational attainment to 6828 (2002). 4. C. A. Rietveld et al., Science 340, 1467 (2013). but these mechanisms are likely to be test whether going further in school has 5. A. Okbay et al., Nature 533, 539 (2016). complex and phenotype-specific. Genetic a causal effect on one’s health. One of the 6. R. de Vlaming et al., PLOS Genet. 13, e1006495 (2017). associations might depend on environ- assumptions of Mendelian randomization 7. E. M. Tucker-Drob et al., Curr. Dir. Psychol. Sci. 22, 349 mental context (6); for example, the al- is that genes are not correlated with unob- (2013). 8. C. Jencks, Am. Sociol. Rev. 45, 723 (1980). leles influencing educational attainment served confounds, such as parental environ- 9. A. L. Price et al., Nat. Genet. 38, 904 (2006). might differ between birth cohorts that ment. The study of Kong et al. clearly shows 10. D. H. Hamer, Mol. Psychiatry 5, 11 (2000). have experienced different social and po- that this assumption can be wrong. If alleles 11. D. A. Lawlor et al., Stat. Med. 27, 1133 (2008). 12. D. J. Benjamin et al., Annu. Rev. Econom. 4, 627 (2012). litical institutions. And genetic effects can identified in genetic association studies are 13. J. M. Horn et al., Behav. Genet. 9, 177 (1979). exert their influence via environmentally not, in fact, independent of the parental en- 14. B. M. D’Onofrio et al., J. Child Psychol. Psychiatry Allied mediated channels, such as how teachers vironment, then Mendelian randomization Discip. 44, 1130 (2003). react to students with specific genotypes studies might detect causal effects that are (7, 8). Now, genetic nurture provides an- not really there. 10.1126/science.aar6429

SCIENCE sciencemag.org 26 JANUARY 2018 • VOL 359 ISSUE 6374 387

Published by AAAS

DA_0126Perspectives.indd 387 1/24/18 10:59 AM Using nature to understand nurture Philipp D. Koellinger and K. Paige Harden

Science 359 (6374), 386-387. DOI: 10.1126/science.aar6429 Downloaded from

ARTICLE TOOLS http://science.sciencemag.org/content/359/6374/386

RELATED http://science.sciencemag.org/content/sci/359/6374/424.full http://science.sciencemag.org/ CONTENT

REFERENCES This article cites 14 articles, 3 of which you can access for free http://science.sciencemag.org/content/359/6374/386#BIBL

PERMISSIONS http://www.sciencemag.org/help/reprints-and-permissions

on February 1, 2018

Use of this article is subject to the Terms of Service

Science (print ISSN 0036-8075; online ISSN 1095-9203) is published by the American Association for the Advancement of Science, 1200 New York Avenue NW, Washington, DC 20005. 2017 © The Authors, some rights reserved; exclusive licensee American Association for the Advancement of Science. No claim to original U.S. Government Works. The title Science is a registered trademark of AAAS.