NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

National Research Council (US) Committee on Advances in Collecting and Utilizing Biological Indicators and Genetic Information in Social Science Surveys; Weinstein M, Vaupel JW, Wachter KW, editors. Biosocial Surveys. Washington (DC): National Academies Press (US); 2008.

Cover of Biosocial Surveys

Biosocial Surveys.

Show details

10Genetic Markers in Social Science Research: Opportunities and Pitfalls

and .

Social science survey research provides rich databases that are informative regarding factors that influence the health and well-being of target populations. These studies frequently are rich in behavioral assessments that can include measures as diverse as cognitive assessment and personality characteristics, social support patterns, health behaviors, and the availability and utilization of health care services. Such research frequently has major advantages, including large, carefully selected samples that are nationally representative or that target particular populations, frequently with long-term longitudinal data to monitor secular changes in risk and protective factors over time or over developmental periods. Examples of such research include such well-known studies as the Health and Retirement Study (HRS, Juster and Suzman, 1995), the National Long Term Care Survey (NLTCS, Research Triangle Institute, 2002), and the National Health and Nutrition Examination Survey (NHANES, Centers for Disease Control and Prevention, 2006).

Recently there have been efforts to incorporate collection of DNA and other biomarkers into these types of surveys through the collection of blood samples or cheek swabs. The incorporation of genetic markers into studies that have health as an emphasis has the potential to provide a more thorough understanding of the mechanisms of the complex factors that influence health outcomes, when considered with the other domains that are frequently assessed, including behavioral, psychological, social, economic, and environmental indicators. Studies such as the MacArthur Study of Successful Aging that have incorporated DNA and biological markers (Crimmins and Seeman, 2001) have demonstrated strong relationships between biological indicators and traditional demographic variables and health outcomes. In the following discussion, issues are considered regarding the incorporation of DNA collection into studies that were designed to address social science questions.


Incorporation of DNA collection into social science surveys has the potential to add considerable value to such studies. The utility of genetic information is based on the observation that there is variability in the allele form of individual genes that can be measured and that can contribute to variability in health-related outcomes. Potential uses include investigation of genes that have well-established influences, such as the Apolipoprotein E (ApoE) gene and its relationship to risk for Alzheimer disease and cardiovascular disease; identification of genes that had not been previously known to have relationships with the outcome variables; investigation of whether an individual's genetic status contributes to variability in the way in which other factors influence outcome variables (gene-environment interaction or gene-gene interaction); investigation of correlations between genetic and environmental factors; and other, more novel uses, such as controlling for genetic status to obtain a more accurate picture of how nongenetic factors are related to the outcome variables. For example, by controlling for genetic status at the ApoE gene, a clearer picture could emerge of the effect of other factors, such as cognitive activity, on the development of dementia.


A central question is whether the complexity of the kinds of outcome variables that are considered in social science survey research is too great to be meaningfully considered in a genetic context. This argument might be considered to be valid if one is limited to the Mendelian perspective, which is the notion that genetic influences on a trait consist of the effect of a single major gene. Much early genetics research was concerned with determining the applicability of the Mendelian rules of transmission to an ever-broadening range of phenotypes. The phenotypes were typically dichotomous (presence or absence of a disease condition is a common type of Mendelian trait), and the principal concern was whether the relative numbers of organisms that were assignable to the different categories conformed to expectations derived from the Mendelian theory. Frequency statistics were adequate to assess the outcomes. If category assignment was reasonably unambiguous, variability in the phenotype within the categories was usually ignorable, along with any influence on that variability due to environmental factors or to genes other than the primary one under investigation. The Mendelian perspective dominated medical genetics through much of the middle of the 20th century, when major progress was made in identifying single major genes that influenced a variety of medical conditions, usually devastating but rare diseases (National Center for Biotechnology Information, 2006).

A parallel avenue of research into heredity was that of biometric or quantitative genetics, which was explicitly concerned with phenotypes that were continuously distributed. The analytical statistics in this approach concerned means, variances, and covariances, and the variance decomposition algorithms provided for a component assignable to heredity, another to environment, and others to correlations and interactions between these agencies. It is this perspective that is most likely to be the more appropriate one for considering the nature of genetic effects on complex traits that are of interest in social science survey studies. It is interesting to note that there has been a substantial emphasis on the analysis of complex traits in the recent genetic literature. This approach is applied widely in psychiatric genetics, behavioral genetics, and genetic epidemiological investigations of complex health-related factors, such as risk factors for cardiovascular disease.

For historical reasons that cannot be detailed here, in certain scientific circles, heredity, under the label of “nature,” came to be regarded as antagonistic or oppositional to environment, labeled “nurture.” The antagonism of these terms was already implied in the presumed source, Shakespeare's “The Tempest,” wherein Caliban is described as “a devil, a born devil, on whose nature nurture can never stick.” Galton (1883) brought the alliteration of nature-nurture into scientific discourse with the same intimation of hostility. Although his perspective was considerably more nuanced, his statement in respect to human faculty that “there is no escape from the conclusion that nature prevails enormously over nurture” was easily translated into a view of nature versus nurture. This perception had significant influence in the behavioral and social sciences, and led to a strong predilection to seek explanations from the environmental realm. Indeed, in some arenas, there was strong a priori rejection of any possibility of genetic involvement. In recent decades, however, a more balanced view has emerged, partaking increasingly of the quantitative genetics perspective attributing main effects to factors from both genetic and environmental domains, and with all of the subtle possibilities arising from the inevitable interaction terms. In its simplest version, the model can conceive of causal influences arising from these separate domains and combining additively to produce the phenotype. But it is increasingly clear that the causal relationships can be much more complex than this, with significant correlations and interactions between nature and nurture. These coactions have substantial implications for the design and interpretation of laboratory and field studies on health-related phenomena (see Moffitt, Caspi, and Rutter, 2005; Plomin and Rutter, 1998).

Plomin, DeFries, and Loehlin (1977) provided a now-classic description and terminology for various types of correlations and interactions in the context of behavioral genetics. Gene-environment correlations occur when the distribution of environmental influences on a particular phenotype is not independent of the distribution of genetic influences on that phenotype. Three classes are distinguished: passive, reactive, and active. A frequently cited example of the passive situation is the provision to children of an intellectually stimulating environment of books, education, and approbation of scholarly pursuits by parents who have also transmitted genes that positively affect cognitive abilities. Reactive correlation arises from situations in which the (genetically influenced) behavior of an individual evokes a particular response from the social environment that tends to support (or inhibit, as the case may be) that behavior. Active correlation ensues when an individual seeks out or alters her or his environment to be consistent with a genetically influenced predilection.

Various sorts of gene-by-environment (G×E) interaction can also be identified (see, for example, Kearsey and Pooni, 1996). In general terms, however, a compact definition can encompass those circumstances in which the effect of a difference in genotype (single gene or polygenic set) depends on features of the environment or, equivalently, in which the effect of differences in features of the environment are dependent on genotype.

The quantitative genetic perspective has been transformed by the wide availability of information about measured polymorphisms from DNA. This can be in the form of measurable polymorphisms in functional genes or as information on massive numbers of polymorphic markers distributed throughout the genome that are not necessarily functional polymorphisms in genes that affect the trait. The incorporation of measured genetic variability has the potential to increase greatly the ability to investigate complex traits, including a genetic perspective, particularly with respect to G×E interactions and other approaches that go beyond the idea of a main effect of an individual gene on a trait.

By virtue of the degree of control that can be exercised in experimental studies with animal models, G×E interactions can be displayed with striking clarity. Because these provide illustrations of the type of interactive phenomena that might also be expected in human studies, we present here some examples to illustrate the phenomena.

Animal Model Studies

Some interactions have emerged from studies in which the environmental differences were in effect for a sustained period of time. Blizard and Randt (1974), for example, reared two inbred strains of mice under one of three housing conditions—standard caging and relatively enriched or impoverished in terms of sensory stimulation. (The animals of an inbred strain are approximately uniform genetically, and different inbred strains differ genetically from each other). The measured phenotype, novel object-oriented activity, was an assessment of exploratory or anxiety behavior exhibited when the animals were placed in a novel environment. Briefly, animals of one strain (C57BL/6) were totally unaffected by differences in rearing environment; animals of the other strain (DBA/2) displayed a strong effect, with animals from the enriched condition displaying twice the level of activity as the impoverished group.

A possible toxic risk factor for development of Alzheimer disease— aluminum in the diet—was explored in a mouse study by Fosmire, Focht, and McClearn (1993). In each of five inbred strains, two groups were established, one being fed a normal control diet and the other a diet enriched in aluminum. The brain aluminum levels of three of the strains were unaltered by diet; one showed a trend toward increase, and one displayed a threefold increase.

Falconer (see Falconer and Mackay, 1996) selectively bred mice for high or low growth from three to six weeks of age on either a standard diet, or on one for which nutritional value had been degraded by dilution with nondigestible fiber. Response to selection occurred in both conditions. There was thus a line of animals that grew large on the good diet and another line that grew less on that diet; another line that grew large on the poor diet and one that grew less on that diet. Testing these groups under the alternate dietary condition made possible an evaluation of the extent to which the same genes influenced growth in the two environments. The genetic correlation was .66, indicating a large “overlap” of the genes involved, but indicating also a very substantial involvement of different genes in the two nutritional environments.

A further example may be drawn from the research literature on aging. Vieira and colleagues (2000) maintained Drosophila under different environmental circumstances throughout their lives. The five environments differed in the temperature of the incubator in which the animals lived—standard control, higher than control, lower than control, a control temperature but with a heat shock administered during pupal stage, and a reduced diet in a standard temperature. Seventeen quantitative trait loci (QTLs) affecting longevity were identified. (QTLs are genes associated with the phenotype but whose location on the chromosomes is known only approximately.) Not one of the QTLs was uniformly influential in all environments. In several cases, the effect was detectable in one environment only. For some, the allele of the QTL that was associated with longer life in one environment was associated with shorter life in another; for some the influence was shown in one sex only; for some, in one particular environment, the “increasing” allele in females was a “decreasing” allele in males. In short, all of the genetic variance was involved in genotype-environment interaction, genetic-sex interaction, or both.

Human Studies

A common feature of these examples is that a presumptive causal element, genetic or environmental, may have differing effect depending on the context of other environmental and genetic elements that are present. Paraphrasing in terms of the health sciences, it may be expected that the virulence of risk factors and the beneficence of remedial or preventive interventions will vary greatly from individual to individual, depending on the unique context presented by each individual. As in the case of the animal model studies, we provide a sampling of the relevant literature, which is large and burgeoning.

Bouchard and colleagues (1990) investigated a putative risk factor— the effects of long-term overfeeding—in young adult male monozygotic twins. For a variety of outcome variables, including gain in body weight, percentage of fat, fat mass, and estimated subcutaneous fat, the variance among pairs was about threefold the variance within pairs. This intrapair similarity is suggestive of involvement of genotype in response to this nutritional environmental intervention (shared environmental influence cannot be ruled out, however).

A polymorphism in the angiotensin-converting enzyme (ACE) that is characterized by the absence (deletion) of a 287-base pair marker has been shown to be related to the response to exercise. In general, the insertion allele (presence of the 287-base pair marker) is associated with greater endurance, whereas the deletion allele is associated with greater muscular strength, although the results have been somewhat conflicting (Folland et al., 2000; Gayagay et al., 1998; Montgomery et al., 1997; Myerson et al., 2001). On the basis of these findings, the relationship of ACE to efficacy of a health-promoting behavior in the elderly was explored by Kritchevsky and colleagues (2005). The risk to septuagenarians of developing mobility limitation was studied as a function of spontaneous activity level. Overall, those more active at baseline were less likely than those who did not exercise to incur limitations to their activity during a follow-up period. However, an interaction was present in that individuals with a particular genotype at this locus (designated II) benefited less than did the other two genotypes (DD or DI).

A well-studied gene with respect to gene-environment interactions related to cardiovascular health is the ApoE gene. This is one of the few genes identified to date that have common variants with well documented and consistent main effects on traditional social science outcomes. Numerous studies have investigated the effect of the ApoE isoform on response to dietary intervention, with inconsistent results. In a meta-analysis of a number of studies, Ordovas et al. (1995) found that the effect of the ApoE-4 showed a higher low-density lipoprotein (LDL) response than other ApoE alleles in diets that reduced total dietary fat, no difference in response in diets that reduced only dietary cholesterol, and a lower LDL response in diets in which the fat saturation (but not total amount of fat) was modified, although not all studies have found this effect (e.g., LeFevre et al., 1997). A differential response of LDL, high-density lipoprotein, and triglyceride levels to diet as a function of the ApoE-2 and ApoE-4 alleles was observed in a Finnish study of children and young adults who were in a free-living situation rather than a dietary intervention study (Lehtimäki et al., 1995). Cardiovascular responsivity to mental stress has also been shown to vary as a function of the ApoE polymorphism (Ravaja et al., 1997). Another ApoE polymorphism in the gene's promoter region (-219G→T) has been shown to affect LDL concentrations differentially (Moreno et al., 2004) and insulin sensitivity (Moreno et al., 2005) in response to a high-fat dietary modification.

Jaffee and colleagues (2005) examined conduct problems in young twins as a function of genetic risk (estimated from co-twin's behavior) and from physical maltreatment. Maltreatment increased probability of a conduct disorder diagnosis by 2 percent in those with the lowest genetic risk and by 24 percent among children with the highest genetic risk.

The intent of this discussion has been to illustrate the potential presence of interactions, not to imply that they are omnipresent. There are clearly variables in both the environmental and genetic domains that have powerful main effects. But interactions may be present, and they can be subtle and powerful. One implication is that identification of a gene as the gene for some attribute must be tentative; it might be without effect or with a different effect in some environment other than that in which it is first described. Similarly, environmental influences may have greater or lesser impact, depending on the genotype of the individuals on whom they impinge. Confidence in generality of the effect will accrue with subsequent observations in altered contexts. An interesting issue is whether an investigation of interaction effects should be undertaken in the absence of observation of main effects. There are methodological issues that arise related to the number of tests for interactions that are possible, resulting in the danger that a blind search for interaction effects in the absence of main effects could lead to a serious problem of false positive results. However, this should not preclude assessment of interaction effects in the absence of main effects under any circumstance. It is possible that subsequent work could provide a line of evidence in favor of a more careful search for interactions without main effects. The example from the meta-analysis of the differential effect of dietary intervention as a function of ApoE allele status noted above is a pertinent example, in which the differential effect of the alleles were in opposite directions, depending on the precise form of the dietary intervention.

In some respects and for some purposes, correlations and interactions such as we have described can be seen as vexing. A more positive view is perhaps warranted. It may well be that the interactions that appear are indications of key processes in the functioning of the complex systems mediating genetic and environmental influences on complex phenotypes. They may be signposts to particularly productive avenues of research.

In summary, a genetic perspective can be incorporated into research despite the likely complexity of genetic and environmental influences and how they interact in complex ways in social science traits. By discarding the nature versus nurture perspective, researchers studying the genetics of complex traits have embraced a more interactive model. The incorporation of measured genetic variability into social science survey research, which frequently has excellent environmental assessment, opens up rich new opportunities to investigate G×E interactions in large samples, with the potential of providing a more realistic description of factors that affect social science survey outcomes.


The potential to investigate more complex, and more realistic, models that incorporate measured genetic information into social science survey studies is a major advantage of obtaining DNA. As noted above, the complexity of the outcomes makes utilizing DNA more challenging than simply including another predictor into a regression model. Researchers involved in genetic studies of complex traits, however, have recognized this complexity and have developed methodologies to deal appropriately with it. In the past, G×E interaction effects could be considered only in the context of latent variable models, which have extremely limited power to detect interactions. The recent explosion in the availability of measurable DNA variability through the use of DNA markers, coupled with advances in the assessment of environmental factors, opens up opportunities to make real progress in understanding these effects on complex traits. There is a danger of improperly dealing with the complexity, resulting in misleading or biased conclusions. It is therefore essential that the use of DNA is rigorously evaluated from a methodological perspective to limit the opportunities for erroneous conclusions to be drawn.

Potential Advantages

In addition to the broad potential advantages in making scientific progress in understanding more completely the nature of genetic and environmental factors on complex social science traits, there are several other practical advantages. One is the potential to have DNA in long-term availability. It is possible to create either immortalized cell lines that provide unlimited access to the DNA indefinitely into the future, or to extract a finite amount of DNA from a blood, cheek cell, or saliva sample for long-term storage. Consequently, there is the potential to have long-term access to DNA for research purposes not yet conceived or for use with technology not yet developed that will make more efficient use of limited samples. Timely collection of DNA can be important in survey research, particularly longitudinal research, in which attrition can result in substantial loss to follow-up of participants in later waves of assessment. For surveys that focus on aging-related issues, there is the more significant issue of loss due to mortality.

At first glance, the issue of G×E interactions seems to introduce an annoying degree of complexity. When an investigator is interested primarily in general characteristics of a population or specific subgroup, interactions do introduce an additional source of variability that can mask the primary effects that are of interest. However, if one takes a different perspective that more closely resembles that of a clinician, the ability to identify individual-level risk factors that are not constant across the group is significant. It has the potential to provide more effective individual-level prediction of outcomes based on risk factors that are unique to the individual.

A related issue is that the identification of individual-level risk factors can be used to develop prospective procedures for identifying individuals who are at high risk for developing adverse outcomes. At a minimum, such individuals can be informed of their elevated risk status so that they can be monitored more closely than standard. Of potentially greater interest is the possibility of targeting interventions specifically to at-risk individuals. The significance of this approach is that while interventions might clearly have an advantage at the overall population level, individuals could respond to an intervention differentially, with its being effective for some, neutral for others, and potentially even damaging to still others.

Potential Disadvantages

The incorporation of DNA collection into social science surveys has several potential disadvantages. At a practical level, there are logistical and financial constraints surrounding the process of sample collection. How severe these constraints are depends on the method of DNA collection and whether it is coupled with assessment of a more extensive battery of biomarker measures. If a survey is primarily conducted by mail or telephone, DNA samples of limited yield can be collected using cheek swabs that are returned through the mail. DNA can be extracted and stored for nominal cost. The drawback is that the amount of DNA is limited, and the failure rate for obtaining a useful DNA sample can be substantial. If blood is drawn, the yield and quality of the DNA improves substantially. However, this requires contact between the participant and a trained professional for the drawing of blood. If the study involves face-to-face contact, the logistical burden can be manageable, but otherwise logistical factors can be challenging. Although the costs of obtaining a DNA sample can be minimal, the cost of genotyping large numbers of individuals from large surveys can be prohibitive unless the genetic component of the study is very tightly focused (for example, limited to genotyping of the ApoE gene).

Another potential disadvantage is related to the participant burden and the purpose of the study. Participants might be recruited to a survey study on the basis of their willingness to be involved in social science research. It might not be clear to participants why they are being asked to provide DNA samples for a study that has been described to them as a social science survey. If participants have been involved longitudinally in a survey, they would not have consented to providing a DNA sample, so additional consent would be required. A request for DNA potentially imposes an additional response burden that participants had not anticipated. If the purpose and procedures for collecting DNA are not thoroughly and properly planned, there is the potential to introduce a bias into the sample that is related to willingness to provide DNA and not related to the primary purpose of the study.

An issue that is not so much a disadvantage but a challenge involves the difficulties that arise when there is a large amount of data generated, in terms of potential type 1 and type 2 error rates and the practical issues that arise in trying to limit the number of statistical tests that are done and account for multiple comparisons. If a study opts to do a full genome scan using single-nucleotide polymorphisms (SNPs) assayed using DNA microchips, there is the potential to generate a million or more DNA markers. Dealing with such massive amounts of data requires specialized techniques in data reduction and haplotype block identification. Substantial methodological work in statistical genetics is currently being conducted to address this issue effectively.


In addition to the more practical issues discussed above, incorporation of DNA collection raises a number of ethical issues (reviewed by Durfy, 2001, in the context of aging-related surveys). Ownership of genetic information is an area that is potentially troublesome, and issues regarding ownership need to be worked out explicitly. Some studies have a policy that requires all information be made publicly available to any investigator who desires access. Other studies involving private biotechnology companies consider the genetic information to be proprietary. It is important for issues of ownership to be clearly planned, and information regarding ownership should be made clear to the participants in the informed consent process.

In addition to access issues, it is important to have a clear plan for the use that will be made of the genetic data, and these issues should be incorporated into the informed consent process. Questions regarding use include whether the use of DNA goes beyond the initial scientific goals of the survey study, whether the data will be made available to other investigators outside the scope of the original project, whether the data will be retained indefinitely so that use in the future cannot be anticipated, and whether results of DNA testing will be provided back to the participant. The last issue is important, since there are few genetic results that are clear-cut risk factors for individuals, particularly in the context of complex traits. In the context of other modifying environmental and genetic factors, individual-level information on even well-known risk factors can be difficult to convey accurately to participants, since risk is generally reported in probabilistic terms. If the decision is to provide information to participants, considerable attention must be paid to how the information is presented. Conversely, if there is clear information about a risk factor, the investigators must consider their obligations to provide this information to participants, either now or if future work clarifies the nature of a genetic risk factor.

While there are important issues regarding privacy and maintaining anonymity with public data sets for any potential identifying information, including demographic factors, these issues merit special attention for genetic data, which contain the risks associated with sensitive medical data in general but also have additional risks. Depending on the extent to which genotyping is undertaken, genetic data can contain sufficient information to identify an individual uniquely. There also is a component of risk related to the familial nature of genetic data in which there is the potential to identify individuals who have not consented to participate in the research but who are related to participants. Finally, misuse of genetic information following breach of anonymity can result in problems associated with stigmatization and discrimination on a variety of factors, such as the availability or cost of health insurance. Because of these concerns, attention is warranted to issues of risk of breach of anonymity in terms of making data publicly available to researchers outside the context of the original study (Annas, Glanz, and Roche, 1995; National Bioethics Advisory Commission, 1999).


The incorporation of DNA into social science surveys can contribute to a more accurate understanding of both genetic and environmental factors that contribute to complex outcomes. Particularly for surveys that focus on health, DNA has a potentially important role to play. Reproductive outcomes, health and diseases, life span and aging, and cognitive function and personality all have substantial genetic influences that are not fully understood. Examples of areas for which genes have been identified include cardiovascular risk factors, diabetes, hypertension, obesity, markers of inflammation, lung function, cognitive function, and addictive behaviors. The extensive environmental assessment information in social science surveys provides a unique opportunity to improve understanding of how these genes function in a broader context, as well as how they contribute to the context in which environmental factors influence social science outcomes.


  1. Annas GJ, Glanz LH, Roche PA. Drafting the Genetic Privacy Act: Science, policy, and practical considerations. Journal of Law, Medicine, and Ethics. 1995;23:360–366. [PubMed: 8715056]
  2. Blizard DA, Randt CT. Genotype interaction with undernutrition and external environment in early life. Nature. 1974;251:705–707. [PubMed: 4427663]
  3. Bouchard C, Tremblay A, Després JP, Nadeau A, Lupien PJ, Thériault G, Dussault J, Moorjani S, Pinault S, Fournier G. The response to long-term overfeeding in identical twins. The New England Journal of Medicine. 1990;322:1477–1482. [PubMed: 2336074]
  4. Centers for Disease Control and Prevention (CDC) National Health and Nutrition Examination Survey data. Hyattsville, MD: U.S. Department of Health and Human Services, Centers for Disease Control and Prevention; 2006. [January 15, 2007]. Available: http://www​​/data/nhanes/nhanes3/nh3gui.pdf.
  5. Crimmins E, Seeman T. National Research Council; Commission on Behavioral and Social Sciences and Education. Integrating biology into demographic research on health and aging (with a focus on the MacArthur Study of Successful Aging) In: Committee on PopulationFinch CE, Vaupel JW, Kinsella K, editors. Cells and surveys: Should biological measures be included in social science research? Washington, DC: National Academy Press; 2001. pp. 9–41.
  6. Durfy SJ. National Research Council; Commission on Behavioral and Social Sciences and Education. Ethical and social issues in incorporating genetic research into survey studies. In: Committee on PopulationFinch CE, Vaupel JW, Kinsella K, editors. Cells and surveys: Should biological measures be included in social science research? Washington, DC: National Academy Press; 2001. pp. 303–328.
  7. Falconer DS, Mackay TFC. Introduction to quantitative genetics. fourth. Essex, England: Longman; Group Ltd.: 1996.
  8. Folland J, Leach B, Little T, Hawker K, Myerson S, Montgomery H, Jones D. Angiotensin-converting enzyme genotype affects the response of human skeletal muscle to functional overload. Experimental Physiology. 2000;85:575–579. [PubMed: 11038409]
  9. Fosmire GJ, Focht SJ, McClearn GE. Genetic influences on tissue deposition of aluminum in mice. Biological Trace Element Research. 1993;37:115–121. [PubMed: 7688525]
  10. Galton F. Inquiries into human faculty and its development. London, England: Macmillan; 1883. p. 241.
  11. Gayagay G, Yu B, Hambly B, Boston T, Hahn A, Celermajer DS, Trent RJ. Elite endurance athletes and the ACE I allele: The role of genes in athletic performance. Human Genetics. 1998;103:48–50. [PubMed: 9737775]
  12. Jaffee SR, Caspi A, Moffitt TE, Dodge KA, Rutter M, Taylor A, Tully LA. Nature X nurture: Genetic vulnerabilities interact with physical maltreatment to promote conduct problems. Development and Psychopathology. 2005;17:67–84. [PMC free article: PMC2768347] [PubMed: 15971760]
  13. Juster FT, Suzman R. An overview of the health and retirement study. The Journal of Human Resources. 1995;30(Suppl. 1995):S7–S56.
  14. Kearsey MJ, Pooni HS. The genetical analysis of quantitative traits. London, England: Chapman and Hall; 1996.
  15. Kritchevsky SB, Nicklas BJ, Visser M, Simonsick EM, Newman AB, Harris TB, Lange EM, Penninx BW, Goodpaster BH, Satterfield S, Colbert LH, Rubin SM, Pahor M. Angiotensin-converting enzyme insertion/deletion genotype, exercise, and physical decline. Journal of the American Medical Association. 2005;294:691–698. [PubMed: 16091571]
  16. LeFevre M, Ginsberg HN, Kris-Etherton PM, Elmer PJ, Stewart PW, Ershow A, Pearson TA, Roheim PS, Ramakrishnan R, Derr J, Gordon DJ, Reed R. ApoE genotype does not predict lipid response to changes in dietary saturated fatty acids in a heterogeneous normolipidemic population. The DELTA Research Group. Dietary Effects on Lipoproteins and Thrombogenic Activity. Arteriosclerosis, Thrombosis, and Vascular Biology. 1997;17:2914–2923. [PubMed: 9409276]
  17. Lehtimäki T, Moilanen T, Porkka K, Åkerblom HK, Rönnemaa T, Räsänen L, Viikari J, Ehnholm C, Nikkari T. Association between serum lipids and apolipo-protein E phenotype is influenced by diet in a population-based sample of free-living children and young adults: The Cardiovascular Risk in Young Finns Study. Journal of Lipid Research. 1995;36:653–661. [PubMed: 7616113]
  18. Moffitt TE, Caspi A, Rutter M. Strategy for investigating interactions between measured genes and measured environments. Archives of General Psychiatry. 2005;62:473–481. [PubMed: 15867100]
  19. Montgomery HE, Clarkson P, Dollery CM, Prasad K, Losi MA, Hemingway H, Statters D, Jubb M, Girvain M, Varnava A, World M, Deanfield J, Talmud P, McEwan JR, McKenna WJ, Humphries S. Association of angiotensin-converting enzyme gene I/D polymorphism with change in left ventricular mass in response to physical training. Circulation. 1997;96:741–747. [PubMed: 9264477]
  20. Moreno JA, Pérez-Jiminéz F, Marín C, Gómez P, Pérez-Martínez P, Moreno R, Bellido C, Fuentes F, López-Miranda J. Apolipoprotein E gene promoter -219G→T polymorphism increases LDL-cholesterol concentrations and susceptibility to oxidation in response to a diet rich in saturated fat. American Journal of Clinical Nutrition. 2004;80:1404–1409. [PubMed: 15531693]
  21. Moreno JA, Pérez-Jiminéz F, Marín C, Pérez-Martínez P, Moreno R, Gómez P, Paniagua JA, Lairon D, López-Miranda J. The apolipoprotein E gene promoter (-219G/T) polymorphism determines insulin sensitivity in response to dietary fat in healthy young adults. Journal of Nutrition. 2005;135:2535–2540. [PubMed: 16251607]
  22. Myerson SG, Montgomery HE, Whittingham M, Jubb M, World MJ, Humphries SE, Pennell DJ. Left ventricular hypertrophy with exercise and ACE gene insertion/deletion polymorphism: A randomized controlled trial with losartan. Circulation. 2001:226–230. [PubMed: 11208681]
  23. National Bioethics Advisory Commission. Research involving human biological materials: Ethical issues and policy guidance. Rockville, MD: National Bioethics Advisory Commission; 1999.
  24. National Center for Biotechnology Information. Online Mendelian Inheritance in Man, OMIM™. 2006. [Sept. 2007]. Available: http://www​
  25. Ordovas JM, Lopez-Miranda J, Mata P, Perez-Jiminez F, Lichtenstein AH, Schaefer EJ. Gene-diet interaction in determining plasma lipid response to dietary intervention. Atherosclerosis. 1995;118:S11–S27. [PubMed: 8821461]
  26. Plomin R, DeFries JC, Loehlin JC. Genotype-environment interaction and correlation in the analysis of human behavior. Psychological Bulletin. 1977;84:302–322. [PubMed: 557211]
  27. Plomin R, Rutter M. Child development, molecular genetics, and what to do with genes once they are found. Child Development. 1998;69:1223–1242. [PubMed: 9768495]
  28. Ravaja N, Räikkönen K, Lyytinen H, Lehtimäki T, Keltikangas-Järvinen L. Apolipoprotein E phenotypes and cardiovascular responses to experimentally induced mental stress in adolescent boys. Journal of Behavioral Medicine. 1997;20:571–587. [PubMed: 9429989]
  29. Research Triangle Institute. 1999 National Long Term Care Survey: Functional and health changes of the elderly Venipuncture, buccal cell, kin, and next-of-kin supplemental studies Final Report: RTI Project 7798. Research Triangle Park, NC: Research Triangle Institute; 2002. [Nov. 2006]. Available: http://www​​.htm.
  30. Rhodes JS, Crabbe JC. Gene expression induced by drugs of abuse. Current Opinion in Pharmacology. 2005;5:26–33. [PubMed: 15661622]
  31. Smith MA, Banerjee S, Gold PW, Glowa J. Induction of c-fos in rat brain by conditioned and unconditioned stressors. Brain Research. 1992;578:135–141. [PubMed: 1511271]
  32. Vieira C, Pasyukova EG, Zeng ZB, Hackett JB, Lyman RF, Mackay T. Genotype-environment interaction for quantitative trait loci affecting lifespan in Drosophila melanogaster. Genetics. 2000;154:213–227. [PMC free article: PMC1460900] [PubMed: 10628982]
Copyright © 2008, National Academy of Sciences.
Bookshelf ID: NBK62439
PubReader format: click here to try


  • PubReader
  • Print View
  • Cite this Page
  • PDF version of this title (7.4M)

Related information

  • PMC
    PubMed Central citations
  • PubMed
    Links to pubmed

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...