Behavioral Genetics and Gene–Environment Interaction (7A) - Psych & Soc

SECTION 1

Historical Context & Motivation

The question of whether human behavior is driven primarily by innate biological factors or by experiential and environmental forces—the classic nature–nurture debate—has animated philosophy, medicine, and psychology for centuries. From Galton's nineteenth-century assertion that genius runs in families to the mid-twentieth-century dominance of strict behaviorism, the pendulum swung between hereditarian and environmentalist extremes. Modern behavioral genetics emerged precisely to move beyond this dichotomy, employing rigorous quantitative designs—twin studies, adoption studies, and family studies—to partition variance in behavioral traits into genetic and environmental components. The field's maturation has revealed that virtually every measurable psychological trait is influenced by both genetic variation and environmental context, and that these two sources of influence are rarely independent of one another.

1869 • Galton's Hereditary Genius

Sir Francis Galton published Hereditary Genius, systematically cataloging the familial clustering of eminence and arguing that intellectual ability is heritable. He later coined the phrase 'nature versus nurture.'

1924 • First Classical Twin Study

Curtis Merriman and Hermann Siemens independently proposed comparing monozygotic (MZ) and dizygotic (DZ) twins to estimate genetic influence on traits, establishing the classical twin design still central to the field.

1979 • Minnesota Study of Twins Reared Apart

Thomas Bouchard launched the landmark MISTRA project, studying MZ and DZ twins separated in infancy. Results showed remarkably high heritability estimates for personality and intelligence, reinvigorating interest in genetic contributions to behavior.

2003 • Caspi et al. — Gene × Environment Interaction

Avshalom Caspi and colleagues demonstrated that a polymorphism in the MAOA gene moderated the effect of childhood maltreatment on antisocial behavior, providing a seminal example of gene–environment interaction (G × E).

2010s • GWAS and Polygenic Scores

Genome-wide association studies (GWAS) began identifying thousands of single-nucleotide polymorphisms (SNPs) of small effect contributing to complex traits, leading to polygenic risk scores and a deeper appreciation of the polygenic architecture underlying behavior.

The central questions that behavioral genetics seeks to address are deceptively straightforward: To what extent do genetic differences among individuals account for observed differences in behavior and psychological traits? How do environments modify or amplify genetic predispositions? And how do organisms actively select and shape the environments they inhabit based on their genotypes? These questions form the conceptual backbone of MCAT Foundational Concept 7A and require a sophisticated understanding of both quantitative genetics and the mechanisms through which genes and environments transact.

SECTION 2

Core Principles & Definitions

Behavioral genetics rests on a set of foundational concepts that bridge molecular biology, psychology, and statistical methodology. At the most fundamental level, it decomposes the observed variation in a trait (the phenotype) into portions attributable to genetic variation (the genotype) and to environmental variation, as well as to interactions and correlations between them. This partitioning is captured by several key conceptual frameworks that every MCAT examinee should master.

Heritability (h²)

The proportion of phenotypic variance in a population attributable to genetic variance. Broad-sense heritability (H²) captures all genetic effects; narrow-sense heritability (h²) captures only additive genetic effects. Heritability is a population-level statistic—it does not indicate the degree to which a trait is 'genetic' in any individual.

Gene–Environment Interaction (G × E)

Occurs when the effect of an environmental exposure on a phenotype differs depending on genotype, or equivalently, when the phenotypic expression of a genotype depends on the environment. The classic example is the differential susceptibility to stress-related disorders based on serotonin transporter polymorphisms.

Gene–Environment Correlation (rGE)

The non-random association between genotypes and environmental exposures. Three subtypes exist: passive rGE (parents provide both genes and environment), evocative rGE (genetically influenced traits evoke environmental responses), and active rGE (individuals select environments matching their genetic predispositions).

Twin & Adoption Study Designs

Twin studies compare MZ (identical, sharing ~100% of segregating DNA) and DZ (fraternal, sharing ~50%) twins to estimate heritability. Adoption studies separate genetic transmission from environmental transmission. Together, these designs form the cornerstone of classical behavioral genetics methodology.

Polygenic Traits

Most behavioral phenotypes are influenced by many genes of small effect rather than single-gene Mendelian inheritance. This polygenic architecture means that genetic risk is distributed continuously, producing normal-distribution phenotypic variation and complicating simple gene-to-behavior mapping.

✦

KEY TAKEAWAY

Think of heritability like the contribution of soil quality to variation in crop yield across a large farm. If every field received identical sunlight and water (uniform environment), then differences in yield would be entirely due to soil quality—heritability would approach 1.0. If all fields had identical soil but different amounts of rainfall, environmental variance would dominate. In reality, both sources vary simultaneously. Heritability quantifies the relative contribution of genetic differences to trait differences in a specific population under specific environmental conditions—change the population or the environment and the estimate changes.

SECTION 3

Visual Explanation: The ACE Model

The ACE model is the standard quantitative framework for decomposing phenotypic variance in twin studies. It partitions total variance into three latent components: A (additive genetic effects), C (shared or common environment), and E (non-shared or unique environment, including measurement error). The diagram below illustrates how MZ and DZ twin pairs differ in their genetic relatedness, which allows the model to estimate each component.

The ACE path model for twin studies. MZ twins share 100% of additive genetic variance (r = 1.0 for A), while DZ twins share only 50% (r = 0.5 for A). Both twin types share 100% of the common environment (C). Non-shared environment (E) is unique to each individual. By comparing the correlation of MZ and DZ twins on a trait, researchers can algebraically solve for A, C, and E.

As illustrated above, the critical inferential leverage comes from the contrast between MZ and DZ twin correlations. If MZ twins are substantially more similar than DZ twins on a trait, this excess similarity is attributable to the additional 50% of shared additive genetic variance that MZ twins possess. The equations at the bottom of the diagram—known as Falconer's formulas—provide a direct algebraic route to estimating each variance component. Importantly, the E component always captures measurement error, so it never equals zero even if there are no true environmental effects.

SECTION 4

Mathematical Framework: Variance Decomposition

The quantitative backbone of behavioral genetics relies on the decomposition of phenotypic variance into genetic and environmental components. Understanding these equations is essential both for interpreting twin study findings and for answering MCAT questions about heritability, concordance, and gene–environment interactions.

PHENOTYPIC VARIANCE DECOMPOSITION

V_P = V_A + V_C + V_E

Where V_P = total phenotypic variance, V_A = additive genetic variance, V_C = shared environmental variance, V_E = non-shared environmental variance (+ measurement error). When standardized so V_P = 1.0, then a² + c² + e² = 1.

FALCONER'S HERITABILITY ESTIMATE

h² = 2 × (r_MZ − r_DZ)

Where r_MZ = intraclass correlation for monozygotic twins, r_DZ = intraclass correlation for dizygotic twins. This formula doubles the difference because DZ twins share half (not zero) of their additive genetic variance.

SHARED ENVIRONMENT ESTIMATE

c² = r_MZ − h²

Since MZ twin correlation reflects both A and C, subtracting the heritability estimate yields the proportion of variance due to the shared environment.

NON-SHARED ENVIRONMENT ESTIMATE

e² = 1 − r_MZ

MZ twins share all genetic and shared environmental variance. Any within-pair differences are therefore attributable to non-shared environment and measurement error.

⚠

Critical Assumption

The classical twin design relies on the equal environments assumption (EEA)—the premise that MZ and DZ twins experience equally similar environments relevant to the trait. Violations of this assumption can inflate heritability estimates if, for example, MZ twins are treated more similarly by parents due to their physical resemblance, and if that differential treatment causally affects the trait under study.

SECTION 5

Gene–Environment Interaction and Correlation

While heritability estimates partition variance into neat components, the reality of gene–environment dynamics is considerably more nuanced. Two critically important phenomena—gene–environment interaction (G × E) and gene–environment correlation (rGE)—describe the ways in which genes and environments are not independent but rather transact dynamically across development. Mastering the distinction between G × E and rGE is a high-yield MCAT objective.

Left panel: Gene × Environment interaction illustrated with the MAOA–maltreatment example. Non-parallel regression lines indicate that the environmental effect on antisocial behavior differs across genotypes. Right panel: Gene–environment correlation, where genotype influences the environments an individual is exposed to via passive, evocative, or active mechanisms.

The left panel of the diagram illustrates G × E interaction using the landmark Caspi et al. (2002) finding: individuals with the low-activity MAOA genotype who experienced severe childhood maltreatment showed markedly elevated antisocial behavior, whereas those with the high-activity genotype were relatively buffered against the same adverse environment. The key visual signature of G × E is non-parallel slopes relating the environmental variable to the outcome across genotype groups. If the slopes were parallel, the genotype and environment effects would be purely additive, and no interaction would be present.

The right panel depicts gene–environment correlation (rGE), which refers to the systematic association between genotype and environmental exposure. In passive rGE, biological parents transmit both genes and an environment congruent with those genes—for instance, musically talented parents both pass on alleles related to musical ability and fill the home with instruments. In evocative (or reactive) rGE, an individual's genetically influenced characteristics elicit responses from others—a temperamentally sociable child may elicit more social engagement from caregivers. In active rGE (niche picking), individuals actively seek out environments compatible with their genetic dispositions—an intellectually curious adolescent gravitates toward libraries and academic clubs. Active rGE becomes increasingly dominant across development as autonomy expands.

SECTION 6

Worked Example: Estimating Heritability from Twin Data

Consider the following scenario: A research team measures extraversion scores in a large sample of twins and obtains an MZ intraclass correlation of 0.52 and a DZ intraclass correlation of 0.23. Using Falconer's formulas, estimate the contributions of additive genetic, shared environmental, and non-shared environmental factors.

Estimating A, C, and E from Twin Correlations

1
Step 1 — Identify Given Values

We are given: r_MZ = 0.52 and r_DZ = 0.23. These are the intraclass correlations for extraversion across monozygotic and dizygotic twin pairs, respectively.

r_MZ = 0.52, r_DZ = 0.23
2
Step 2 — Calculate Heritability (h²)

Apply Falconer's formula: h² = 2 × (r_MZ − r_DZ) = 2 × (0.52 − 0.23) = 2 × 0.29 = 0.58. This indicates that approximately 58% of the phenotypic variance in extraversion in this population is attributable to additive genetic factors.

h² = 0.58 (58% additive genetic variance)
3
Step 3 — Calculate Shared Environment (c²)

The shared environment component is: c² = r_MZ − h² = 0.52 − 0.58 = −0.06. A negative value is theoretically impossible and suggests that c² is essentially zero. In practice, small negative values arise from sampling fluctuation and are typically rounded to zero, indicating negligible shared environmental influence on extraversion in this sample.

c² ≈ 0 (shared environment negligible)
4
Step 4 — Calculate Non-Shared Environment (e²)

The non-shared environment component is: e² = 1 − r_MZ = 1 − 0.52 = 0.48. This component captures both true non-shared environmental effects (unique experiences that make twins different from each other) and measurement error in the extraversion assessment.

e² = 0.48 (48% non-shared environment + error)
5
Step 5 — Interpret and Verify

Summing: h² + c² + e² ≈ 0.58 + 0 + 0.48 = 1.06. The slight overshoot (from the negative c² being rounded to zero) reflects sampling error. The key interpretation: extraversion in this population is substantially heritable, with the remaining variance attributable primarily to unique experiences and measurement noise rather than shared family environment. This pattern—high h², near-zero c², moderate e²—is actually quite typical for personality traits in adulthood and aligns with decades of behavioral genetics research.

Extraversion: ~58% genetic, ~0% shared environment, ~42% unique environment

SECTION 7

Strengths, Limitations, and Methodological Comparisons

No single research design in behavioral genetics is without limitations. Understanding the strengths and weaknesses of each major design is essential for critically evaluating research findings and for the MCAT's emphasis on scientific reasoning and research design literacy. The table below compares the three cornerstone methodologies.

Design	Strengths	Key Limitations
Twin Studies	Powerful decomposition of A, C, E; large registries available; applicable to any measurable trait; can be extended to multivariate models	Equal environments assumption may be violated; cannot distinguish additive from non-additive genetic effects without additional data; twins may not represent the general population (e.g., prenatal environment differences)
Adoption Studies	Cleanly separates genetic from environmental transmission; biological vs. adoptive parent comparison is conceptually straightforward; can estimate passive rGE	Selective placement (adoptive homes matched to biological family characteristics) confounds estimates; sample sizes often small; adoptive families may have restricted range of environments; prenatal environment shared with biological mother
Family Studies	Easy to conduct; can estimate familial aggregation; applicable to rare conditions; provides concordance rates for clinical genetics	Cannot disentangle genetic from shared environmental effects because family members share both; no estimate of heritability without additional assumptions; confounded by assortative mating
GWAS / Molecular	Identifies specific genetic variants; no family data needed; generates polygenic scores for prediction; can test G × E with measured genotypes	'Missing heritability' problem: GWAS-identified variants explain only a fraction of twin-study heritability; population stratification can confound results; very large samples required; limited to common variants

Comparison of Major Behavioral Genetics Research Designs

✦

KEY TAKEAWAY

Consider each behavioral genetics design as analogous to a different lens on a compound microscope. Twin studies provide a wide-angle view of the relative contribution of genes versus environments, while GWAS offers a high-magnification view of individual molecular variants. Neither lens alone gives the full picture—the most informative research programs triangulate across multiple designs to converge on robust conclusions. On the MCAT, watch for questions that ask you to identify the assumption being violated or the limitation of a particular study design.

SECTION 8

Connections to Epigenetics and the Diathesis–Stress Model

Behavioral genetics provides the statistical architecture for understanding trait variation, but contemporary research has extended these concepts in two particularly important directions that are relevant to the MCAT. First, epigenetics reveals that environmental exposures can modify gene expression without altering the DNA sequence itself—through mechanisms such as DNA methylation and histone modification—providing a molecular substrate for G × E interactions. Second, clinical psychology has formalized the interaction concept within the diathesis–stress model and its more recent extension, the differential susceptibility model.

Feature	Diathesis–Stress Model	Differential Susceptibility Model
Core Claim	Genetic vulnerability ('diathesis') leads to pathology only when activated by environmental stress	Certain genotypes confer heightened sensitivity to environmental influence 'for better and for worse'—both risk in adversity and benefit in enrichment
G × E Pattern	Cross-over absent; 'vulnerable' genotype performs worse under stress, comparable under no stress	Cross-over present; 'susceptible' genotype performs worst under adversity but best under supportive conditions
Genotype Framing	'Vulnerability' alleles; purely negative connotation	'Plasticity' alleles; neutral connotation—high sensitivity to all environments
Example	5-HTTLPR short allele → increased depression only after stressful life events	5-HTTLPR short allele → increased depression after stress but also increased well-being in supportive environments (Belsky & Pluess, 2009)
Clinical Implication	Prevention focuses on reducing stress exposure for vulnerable individuals	Intervention enrichment may be especially effective for susceptible individuals

Diathesis–Stress vs. Differential Susceptibility Models

The MCAT may present scenarios in which you must differentiate between these models. The critical distinction is whether the 'at-risk' genotype shows a cross-over effect: if individuals with the susceptibility allele perform worse than comparison genotypes under adverse conditions and better under enriched conditions, the data support the differential susceptibility model. If the susceptibility allele is only associated with worse outcomes under stress (but equivalent outcomes otherwise), the diathesis–stress model is more appropriate. Both models are instances of G × E interaction but carry different implications for intervention and for our understanding of genetic 'risk.'

ℹ

Epigenetic Mechanisms and G × E

Epigenetic research provides a molecular mechanism for how environmental exposures get 'under the skin.' For example, Meaney and colleagues demonstrated that maternal licking and grooming behavior in rats alters DNA methylation patterns at the glucocorticoid receptor gene promoter in offspring hippocampus, persistently changing stress reactivity. This shows how a purely environmental input (maternal behavior) produces lasting changes in gene expression—without altering the DNA sequence—linking behavioral genetics to molecular biology in a mechanistically satisfying way.

SECTION 9

Practice Problems

Question 1: CONCEPTUAL

A researcher reports that the heritability of intelligence is 0.80 in a sample of upper-middle-class families. A colleague argues that this means 80% of an individual's intelligence is determined by genes. Is the colleague's interpretation correct? Explain why or why not, and describe how heritability might differ in a sample from lower socioeconomic backgrounds.

Show answer

The colleague's interpretation is incorrect. Heritability (h²) is a population-level statistic that quantifies the proportion of variance in a trait attributable to genetic differences among individuals within a specific population and environment. It does not describe the degree to which genes determine the trait in any single person. In upper-middle-class families, environmental variation relevant to cognitive development (nutrition, educational access, stimulation) may be restricted, inflating the relative contribution of genetic variance and yielding a high h². In a more socioeconomically diverse sample, where environmental variation is greater, the same trait would likely show lower heritability because environmental factors now account for a larger share of the total variance. This illustrates that heritability is context-dependent and cannot be extrapolated from one population to another.

Question 2: BASIC CALCULATION

In a twin study of depressive symptomatology, the MZ intraclass correlation is 0.68 and the DZ intraclass correlation is 0.34. Using Falconer's formulas, calculate h², c², and e².

Show answer

Applying Falconer's formulas: h² = 2 × (r_MZ − r_DZ) = 2 × (0.68 − 0.34) = 2 × 0.34 = 0.68. The shared environment component is c² = r_MZ − h² = 0.68 − 0.68 = 0.00. The non-shared environment component is e² = 1 − r_MZ = 1 − 0.68 = 0.32. Notice that the MZ correlation is exactly double the DZ correlation (0.68 = 2 × 0.34), which is the expected pattern when c² = 0 and all familial resemblance is due to additive genetic effects.

Question 3: INTERMEDIATE

A child with a genetic predisposition toward high activity levels is placed in an adoptive home. The adoptive parents, noting the child's energy, enroll the child in sports programs and outdoor activities. Identify the type(s) of gene–environment correlation at play and explain how this dynamic might inflate heritability estimates in a standard adoption study.

Show answer

This scenario illustrates evocative gene–environment correlation: the child's genetically influenced high activity level evokes a specific environmental response from the adoptive parents (enrolling in sports). As the child grows and actively seeks out physical activities on their own, active gene–environment correlation (niche picking) also becomes relevant. In a standard adoption design, the correlation between the child's trait and the adoptive family environment would typically be attributed to 'environmental' influence, since the adoptive parents are genetically unrelated to the child. However, because the environment (sports programs) was itself partly caused by the child's genotype, what appears to be an environmental effect is actually genetically mediated. This can lead to an underestimation of genetic influence in adoption studies (or, from the other direction, inflate the apparent effect of the adoptive environment) when rGE is not modeled explicitly.

Question 4: APPLIED

A clinical trial tests whether a cognitive-behavioral therapy (CBT) intervention reduces anxiety in adolescents. The researchers genotype participants for a polymorphism in the serotonin transporter gene (5-HTTLPR). They find that adolescents with the short/short genotype show dramatically greater anxiety reduction from CBT compared to those with the long/long genotype, who show modest improvement regardless of treatment condition. What type of gene–environment interaction model best explains these findings, and what are the clinical implications?

Show answer

These findings are best explained by the differential susceptibility model. The short/short genotype carriers demonstrate heightened sensitivity to the environment 'for better and for worse': they likely experienced greater anxiety in the absence of intervention (consistent with vulnerability) but also benefited most dramatically from the positive environmental input of CBT. The long/long genotype carriers, by contrast, showed relatively stable, moderate responses regardless of environmental condition—they are less malleable. The clinical implication is substantial: rather than viewing the short/short genotype as a purely 'risk' allele (as the diathesis–stress model would suggest), it functions as a 'plasticity' allele. This means that treatment resources might be most efficiently deployed toward these genetically susceptible individuals, who stand to gain the most from enriched therapeutic environments.

Question 5: CRITICAL THINKING

A meta-analysis of twin studies reports that the heritability of antisocial behavior is h² = 0.50, c² = 0.20, and e² = 0.30. Separately, a GWAS identifies specific genetic variants that collectively explain only 5% of the variance in antisocial behavior. Discuss at least three reasons for this 'missing heritability' gap and explain how gene–environment interaction might contribute to the discrepancy between twin-based and molecular-based heritability estimates.

Show answer

The 'missing heritability' gap—the discrepancy between the 50% heritability from twin studies and the 5% from GWAS—can be attributed to several factors. (1) Polygenicity with small effects: antisocial behavior likely involves thousands of variants, each contributing a tiny proportion of variance; current GWAS may lack the statistical power to detect all of them. (2) Rare variants: standard GWAS arrays capture common SNPs (MAF > 1%) but miss rare variants that may have larger effects. (3) Non-additive genetic effects: epistatic (gene–gene) interactions and dominance effects contribute to twin-study heritability but are largely invisible to standard GWAS additive models. (4) Gene–environment interaction: twin-based h² captures genetic effects that may be conditional on environmental context, inflating genetic variance estimates in the twin design. GWAS, which typically estimates main effects of alleles across diverse environments, may average out these context-dependent effects, yielding lower estimates. (5) Gene–environment correlation: in twin studies, genetic effects that operate through genetically correlated environments are absorbed into the A component, but in GWAS, these indirect pathways are not captured by the measured SNP effects alone. Together, these factors illustrate that twin heritability and GWAS heritability answer somewhat different questions, and the gap is an active area of research rather than a flaw in either approach.

SUMMARY

Summary: Behavioral Genetics and Gene–Environment Interaction

Summary

Behavioral genetics decomposes phenotypic variation into additive genetic (A), shared environmental (C), and non-shared environmental (E) components using the ACE model. Heritability (h²) is a population-level statistic estimated via Falconer's formula: h² = 2(r_MZ − r_DZ), and it is context-dependent—changing with the population and environment sampled. The classical twin study, adoption study, and family study designs each offer distinct advantages and assumptions, and converging evidence across designs strengthens conclusions.

Beyond simple variance partitioning, gene–environment interaction (G × E) describes how the phenotypic impact of an environment depends on genotype (and vice versa), while gene–environment correlation (rGE) describes the non-random association between genotypes and environments through passive, evocative, and active mechanisms. The diathesis–stress model and the differential susceptibility model formalize G × E in clinical contexts, with the latter predicting cross-over effects where 'plasticity' alleles confer both heightened risk in adversity and heightened benefit in enrichment. Epigenetic mechanisms such as DNA methylation provide the molecular basis for environmental effects on gene expression, bridging the gap between behavioral phenotypes and molecular biology.