Genes load the gun, environment pulls the trigger: nature and nurture reframed

The Short Version

Across four very different conversations, on obesity, autism, human individuality and sexual development, the experts land on the same picture: genes rarely fix a trait outright; they set a predisposition, and the environment (in the broadest sense, starting in the womb) decides whether and how much that predisposition shows up. Neuroscientist Zachary Knight puts it in the field's own shorthand: "genetics loads the gun and environment pulls the trigger" [1]. Karen Parker describes autism as a largely polygenic condition whose environmental risk factors may drive it in one child and not another [2], and David Linden argues that "nature versus nurture" should become "heritability interacting with experience filtered through the random nature of development" [3]. Andrew Huberman's walk through sexual development shows the same logic in hormones: the genetic program only works through hormones and receptors, and outside chemicals can push on those pathways [4]. The shared warning is that heritability is widely misread: a trait can be 80% heritable and still change dramatically across a population, and a heritability number belongs to a particular population in a particular environment, not to the trait itself [1][3].

Table of contents

What They Said

Zachary Knight: body weight is highly heritable, and obesity still exploded

Knight, a UCSF physiologist who studies hunger, starts from a number most people don't expect. Comparing identical and fraternal twins, the heritability of body weight comes out at "something on the order of 80%" [1]. He adds that "one of the only things we know about that's actually more heritable than body weight is height," and that "Most diseases are not as heritable as body weight" [1]. Huberman's reaction is telling: "So I don't think most people appreciate that" [1].

The genes involved are mostly brain genes. Among people with severe obesity from early childhood, "something on the order of 10%" have mutations in one appetite pathway, the POMC and melanocortin-4 receptor system, with the receptor the most common site [1]. But most obesity is polygenic: "Something on the order of 1,000 genes have been linked to bodyweight regulation," and "the vast majority of those are expressed in the brain" [1]. That fits, Knight says, because body weight is driven by food intake, and "the brain controls behavior, and also the brain controls energy expenditure" [1]. The classic single-gene cases show how powerful a genetic lesion can be: the ob mouse, which cannot make the fat hormone leptin, and the db mouse, which lacks its receptor, are about three times normal size and eat constantly [1].

Then comes the puzzle Knight sees many people struggle with. Obesity has soared since roughly the 1970s, yet human genetics "has not changed in that amount of time. It's just not fast enough for people to evolve" [1]. How can a trait be 80% genetic and still change so fast? His answer is the core of this whole insight:

• In any society at any moment, people spread out along a distribution of body weight, and "where you lie on that distribution is determined primarily by genetics" [1]. You may be "the person who has the thrifty genes to basically cause you to save energy," or someone whose genes make you "a little bit less hungry" [1].
• The environment then "shifts that whole distribution," so the average moves and most people get heavier [1].

Hence the phrase "genetics loads the gun and environment pulls the trigger" [1]. Some people carry latent sensitivities to palatable food; "in an earlier time they may have been lean," but in today's world of ultra-processed, highly palatable food they gain a great deal of weight [1]. The reverse also holds: anyone can be made lean in a restricted environment. In the World War II Minnesota Starvation Experiment, volunteers fed about 1,600 calories a day lost huge amounts of weight, grew cold, slowed their heart rates and became obsessed with food [1].

Knight is careful about which environmental change matters. Huberman lists popular theories, "everything from seed oils to snacking to smartphones," and suggests cheap, ubiquitous calories, noting a vending machine or restaurant "every 20 meters or so" in an airport [1]. Knight calls that "a plausible hypothesis" and says it "would be surprising to me if it didn't contribute," but warns that "these population level questions is just so hard to actually know because you can't do an experiment" [1]. The best controlled evidence he offers is Kevin Hall's inpatient study at the NIH: people ate far more of ultra-processed meals than of minimally processed meals rated equally palatable, gained weight on them, and lost it when switched [1].

He also shows how the genetic setpoint pushes back once the environment is changed by an individual. After weight loss, leptin falls and hunger neurons become more active; every kilogram lost lowers daily energy expenditure by about 30 kilocalories, and in Hall's covert weight-loss study appetite rose by about 100 calories a day for every two pounds or so lost [1]. Rudy Leibel found that people who had lost around 100 pounds burned about 25% less energy than never-obese people of the same size, for at least a year [1]. Knight concludes that rising hunger is the main reason lost weight returns [1]. Huberman, for his part, calls the willpower-versus-pharmacology debate silly, since people can use any safe combination of tools [1].

Karen Parker: autism as an inherited, polygenic liability shaped by environment

Parker, who directs Stanford's Social Neurosciences Research Program, describes autism in the same terms. "Autism is highly heritable, so there's about 40 to 80 percent of autism is genetic," she says, adding that the estimates "vary wildly" [2]. The common view is that "about 50 percent of autism is associated with common genetic variance," so "autism is largely an inherited polygenic condition" [2]. Single high-impact genes exist too, and can produce severe intellectual disability [2].

Like Knight's body-weight distribution, autistic traits are "continuously distributed across the general human population" [2]. She calls the underlying genetic contribution a "genetic liability," which she explains as "a genetic risk that underlies this continuum of behavioral traits" [2]. People without a diagnosis in intense STEM fields such as engineering, physics and math tend to score higher on measures like the Autism Quotient [2]. Huberman, framing his question to her, states the principle directly: "It's never nature or nurture. It's always an interaction," and not only through the epigenome, because "our genome impacts the way that we interact with the environment" [2].

On the environmental side, Parker lists known risk factors: "advanced parental age," severe prematurity and "Maternal illness during pregnancy" [2]. Her key point is about how genes and environment interact differently in different children. Because the population is so genetically varied, "an environmental risk factor might be a driver for one kid, but not another" [2]. She compares the field to "being in a dark room, and you have a flashlight and you only see where you shine the light" [2]. Her proposed fix is to study "genetically defined subgroups" (fragile X, Prader-Willi, Timothy syndrome) and test "gene by environment interactions" and treatments within each, much as cancer care moved to molecular profiling [2]. That diversity is summed up in her line: "if you've met one kid with autism, you've met one kid with autism" [2]. Huberman goes further: "You could imagine that there are 50 or 60 different spectrums," a space with "multiple dimensions" rather than one line from severe to mild [2].

Her own research shows a biological predisposition that appears early. In rhesus macaques, low vasopressin in spinal fluid marks the less social animals; in children, lower spinal fluid vasopressin tracks greater social symptom severity; and infants later diagnosed with autism already had low levels, suggesting low vasopressin comes before, not after, years of social difficulty [2]. A small trial of vasopressin nasal spray improved social abilities, with a larger trial underway [2]. Her graduate work on voles adds a striking case of latent capacity: meadow voles, normally non-monogamous, helped raise pups under short winter-like days, and when she infused vasopressin into males' brains it was like turning on a light switch; Huberman notes this shows latent circuits can be unmasked rather than built from scratch [2].

She is equally clear about a proposed environmental trigger that doesn't hold up. Wakefield's vaccine paper was retracted over fabricated data, and "every single study that I'm aware of does not show a relationship between vaccination and autism" [2]. She regrets that fear around the topic kept researchers away from studying immune dysregulation that some parents report, which deserves evidence-based investigation [2].

David Linden: genes, experience and chance

Linden, a Johns Hopkins neuroscientist and author of Unique, gives the most systematic account of what heritability means. Very few traits are entirely heritable or entirely unheritable; "most fall in between" [3].

At one end is earwax. Wet versus dry earwax is set by a single gene, ABCC11: "It doesn't matter what foods you ate growing up. It doesn't matter what what diseases your mother had when you were on the womb. It's 100%. Heritable" [3]. Even so, he warns against calling it the earwax gene: "Genes don't code for traits. They code for proteins" [3]. The same gene works all over the body, and "the wet earwax variant of the ABCC 11 gene also confers a slightly higher risk for breast cancer" [3]. At the other end is speech accent, 0% heritable and learned mostly from childhood peers [3].

The middle is where the misunderstandings live. Height is about 85% heritable in the United States, but "If you go to a place where people routinely don't get enough nutrition and are routinely fighting off infectious diseases," as in rural Bolivia or rural India, "It's only 50% heritable" [3]. The reason is that people there "can't live up to their genetic potential for height" [3]. IQ follows the same pattern: roughly 60 to 70% heritable in affluent populations and less among people facing poverty or discrimination [3]. His conclusion is a moral one: "everyone needs to have basic things like the ability to learn and enough nutrition and decent medical care and schools in order to fulfill their genetic potential" [3].

He describes how heritability is measured (comparing identical and fraternal twins, which assumes families provide equal environments, or better, twins reared apart) and what the Minnesota Study of Twins Reared Apart found: the Big Five personality traits are roughly 50% heritable, yet family upbringing explains almost none of the rest [3]. Parents can pass on values, religion or occupations, but not these core traits [3].

The rest comes largely from chance. The genome is, in his words, "a rather vague recipe" [3]. It might tell a group of neurons that roughly half should cross the brain's midline, so "for some individuals, 40% of their axons will cross the midline of the brain. And for another individual 60% will even in identical twins" [3]. Newborn identical twins can differ by 30% in the size of a spleen or liver, "Even though they have the exact same DNA," because of "the random or as we say stochastic nature of neural development" [3]. Nine-banded armadillos, born as identical quadruplets, differ early in brain wiring and boldness [3]. And these chance traits aren't inherited: "if you, through the randomness of development, happened to have like a particularly great liver, you're not going to pass that on to your children" [3]. Huberman adds that "going to the gym and getting fit does not make your children more fit," because the germ line is not directly modified [3].

Linden also shows traits that look genetic but are set by early experience, "just the sort of trait that if you asked someone they would probably guess is heritable, but actually is not" [3]:

• Nearsightedness. Children who spend less time outdoors in roughly their first five years are much more likely to become nearsighted, because light appears to trigger growth factors that control eyeball elongation [3].
• Heat tolerance. In World War II, Japanese soldiers from cold Hokkaido suffered heat stroke far more than those from semi-tropical Kyushu. Both had the same number of sweat glands, but southerners had more of them innervated, and northern children raised in the south developed the southern pattern, and vice versa [3].
• Vole fur. Voles gestated under lengthening days are born with thin summer fur, and those under shortening days with thick winter fur [3].
• Perfect pitch. Twin studies suggest heritability of roughly 30 to 40%, and early ear training greatly increases the odds of developing it [3].

Genes shape perception too. For the odor androstenone, some people smell nothing, some smell cut grass, and some smell urine or sweat, and "it just depends on genetic variation in one particular odorant receptor" [3]. Yet most smell preferences are learned: babies must learn that poop is disgusting, and whether vanilla smells sweet depends on whether a culture pairs it with sugar [3].

He adds an important boundary on the environment side: single-generation effects in pregnancy are well established, but transgenerational epigenetic inheritance in mammals is not. Male children in utero during the 1918 flu pandemic were a millimeter or two shorter on average, and schizophrenia rates rose from about 1% to 4%, with what we'd now call autism also rising [3]. Gloria Choi's lab at MIT traced a mouse version to the immune molecule interleukin-17, which disorganizes the developing cortex; blocking it prevented the effect [3]. But the human claims of inheriting a grandparent's famine or trauma through the germ line lack a mechanism and often skipped statistical corrections; you may inherit a grandmother's trauma, he says, but probably socially [3].

Finally, his reframing. He has "a lot of problems with nature versus nurture as an expression," popularized by Francis Galton [3]. Nurture is too narrow: "it should be replaced with the word experience," meaning "anything that impinges on you starting from the earliest stages of fetal developments, continuing to the last day of your life," including the foods your mother ate, the diseases she fought off and "the bacterial population of your guts" [3]. And "versus" is wrong because genes and experience interact. The genetic disease PKU only matters if you eat phenylalanine: "If you don't, it doesn't matter that you inherited these things [3]. With athletic ability, a child born fast is more likely to practice and get faster, so genes and experience are feeding back on each other in a positive feedback loop [3]. His replacement phrase: heritability interacting with experience filtered through the random nature of development" [3].

Andrew Huberman on hormones: a genetic program run through chemicals the environment can reach

In his solo episode on sexual development, Huberman traces the chain from chromosomal sex to gonadal, hormonal and morphological sex, asking listeners to "discard with the cultural connotations" and focus only on biology [4]. Genes on the Y chromosome encode Müllerian inhibiting hormone and, through genes like SRY, "promote the formation of testes" [4]. From there, hormones carry the program: sex steroids are lipophilic, pass into cells and "interact with the DNA of cells in order to control gene expression," changing what cells become "in a long-term way" [4].

Several of his examples show that a gene's effect depends on the rest of the chain:

• The Guevedoces. In a mutation first identified in the Dominican Republic, children lack 5-alpha reductase, the enzyme that makes DHT from testosterone. They are born looking female and are often raised as girls, then around 11 to 13 pubertal testosterone causes a penis to grow [4].
• Androgen insensitivity syndrome. Individuals are XY with internal testes that make testosterone, but their androgen receptor is mutated, so "their body can't make use of the testosterone because they don't have the receptors"; they look completely female and generally report feeling like girls and later women [4]. The principle: a hormone must bind a functioning receptor to have any effect [4].
• Estrogen masculinizes the brain. Neurons convert testosterone to estrogen with the enzyme aromatase, and "the masculinization of the brain is not accomplished by testosterone. It is accomplished by estrogen" [4]. As he puts it, "nature is far more interesting than that" [4].
• Beards and baldness. The density of DHT receptors on the face and scalp is genetically determined and dictates beard density and baldness patterns, which is why some men with high testosterone and DHT have light beards [4].

He then turns to the environmental side: "there are things that are environmental, and there are things that people use that actually can impact hormone levels and can impact sexual development in fairly profound ways" [4], sometimes "regardless of chromosomal background" [4]. Tyrone Hayes linked the herbicide atrazine to severe testicular malformations in frogs, with 10 to 92% of males affected at some U.S. sites [4]. Herbicide concerns once dismissed as "hippy-dippy stuff" now rest on "very solid data from federally funded labs," Huberman says [4]. In humans, average sperm density in the U.S. and Western Europe fell from 113 million per milliliter in 1940 to 66 million in 1990 [4]. He reports that "cannabis promotes estrogenic activity by increasing aromatase," which may explain higher rates of gynecomastia in male users, and that beer and other grain alcohols can increase estrogenic activity, a concern during puberty [4]. And he stresses that "puberty has a beginning, a middle and an end, but development is really our entire lifespan" [4].

Where They Agree, and Where They Don't

Agreement on the core model. All four describe genes as setting a propensity rather than a fate. Knight's distribution model (genes decide where you sit, environment moves everyone) [1], Parker's "genetic liability" underlying a continuum of traits [2], Linden's heritability "interacting with experience" [3], and Huberman's gene-to-hormone-to-receptor chain that outside chemicals can disrupt [4] are versions of one idea. Huberman says it in two of the conversations: "It's never nature or nurture. It's always an interaction" [2].

Agreement that heritability is misunderstood. Knight's 80% heritable trait that still exploded in two generations [1] and Linden's height that is 85% heritable in the U.S. but 50% in rural Bolivia [3] make the same point from opposite sides: heritability describes variation among people within a given environment, not how much a trait can change.

Agreement on latent capacity. Knight's people who "may have been lean" in an earlier environment [1], Parker's meadow voles whose fathering was switched on by vasopressin [2], and Linden's Hokkaido children who developed southern sweat glands when raised in the south [3] all show capacities that exist in the biology and appear only under the right conditions.

Different emphases. Knight and Parker speak mostly in two terms, genes and environment. Linden insists on a third: random development, which explains why identical twins differ and which isn't inherited [3]. Linden's definition of environment is also wider than most people's, reaching back into the womb and into the gut [3], which matches Parker's prenatal risk factors [2] and Huberman's focus on pregnancy and puberty [4].

Different levels of certainty about environmental triggers. Knight refuses to name the obesity trigger with confidence, because population-level causes can't be tested experimentally [1]. Linden is sharply skeptical of transgenerational epigenetic claims in mammals [3]. Parker rejects vaccines as a cause on the evidence [2]. Huberman, by contrast, is more willing to act on suggestive data: he attributes falling sperm counts largely to widely used herbicides, and on cell phones cites a 2013 rat study and later studies of varying quality [4]. That gap is real: the guests tend to hold environmental claims to the standard of controlled evidence, while Huberman's hormone episode leans toward precaution where data are thinner, and says so.

Why It Works

The sources give several mechanisms for how a predisposition becomes a trait.

Genes build proteins, not traits. Linden's point that genes "code for proteins" [3] explains why the same gene can matter in one setting and not another. Knight's body-weight genes are mostly brain genes that shape hunger and energy expenditure [1], so their effect depends on what food is around. The PKU gene only causes harm if phenylalanine is eaten [3].

Distributions versus averages. Heritability measures how much of the variation between people traces to genes. When the environment changes for everyone, the whole curve shifts while people's rank order can stay largely genetic [1]. When the environment is poor, differences in nutrition and disease swamp genetic differences, and heritability drops [3].

Hormones and receptors as the bridge. Sex steroids enter cells and change gene expression long-term [4]. But a hormone needs its receptor (androgen insensitivity), an enzyme may need to convert it (5-alpha reductase, aromatase), and receptor density can be genetic (DHT receptors on the scalp) [4]. Every link is a place where genes and environment meet: cannabis raises aromatase activity, and herbicides alter hormone ratios [4].

Homeostasis defends the setpoint. In obesity, genetically tuned circuits push back against change. Falling leptin after weight loss activates hunger neurons, energy expenditure falls, and appetite rises [1]. Knight and Huberman agree, "It's hard to beat homeostasis" [1].

Critical windows. Early life is when experience has the most leverage: light exposure in the first five years sets eyeball growth [3], heat in childhood sets sweat gland innervation [3], a maternal infection can disorganize cortex through IL-17 at a sensitive window [3], and in Parker's data low vasopressin is already present in infancy [2].

Stochastic development. Because the genome is a vague recipe, development involves chance, producing differences even between identical twins [3].

Where It Applies, and Where It Doesn't

The model applies to polygenic, continuously distributed traits: body weight [1], autistic traits [2], height, IQ and personality [3]. For these, "80% heritable" does not mean "80% fixed," and environmental changes can shift outcomes a lot for a whole population [1][3].

It applies less to a few single-gene traits. Earwax type is 100% heritable no matter what you eat or how you're raised [3]. Some single-gene mutations in the melanocortin pathway cause severe obesity from childhood [1], and androgen insensitivity or 5-alpha reductase deficiency determine major features of development [4]. Even here the gene works through a pathway (a hormone, a receptor), which is why Knight notes that the one approved melanocortin drug is used mainly for people with mutations in that pathway [1].

It does not mean every environmental claim is true. Knight warns that population-level explanations are hard to test [1], Parker describes how the fraudulent vaccine theory wasted money [2], and Linden says transgenerational epigenetic inheritance in humans lacks evidence [3]. The environment matters, but specific triggers need evidence.

It also doesn't mean individuals can easily override their predisposition. Knight is candid that diets designed to keep weight off have been very unsuccessful, and that rising hunger is the main obstacle [1]. The environment is powerful at the population level and early in life; for an adult fighting a defended setpoint, the fight is real, which is why he sees medication as a legitimate tool [1].

Finally, heritability is not portable. A number from affluent American twins doesn't describe people in rural Bolivia or India [3], and twin-based estimates rest on assumptions, such as families providing equal environments [3].

For anything touching weight management, autism, hormones or pregnancy, talk to a qualified doctor before acting on these ideas.

What To Do With It

1. Read heritability numbers as population statements, not personal destiny. An 80% heritable trait can still change sharply when the environment changes, as obesity did in two generations [1], and a trait's heritability rises or falls with living conditions [3].

2. Shape the food environment rather than relying on willpower alone. Knight's grounded advice is to eat fewer ultra-processed foods and more whole foods, get enough protein, and drink water so you aren't eating when you're thirsty [1]. Kevin Hall's study shows the food environment itself changes how much people eat [1].

3. If you've lost weight, expect your biology to push back, and plan for it. Appetite rises by about 100 calories a day per two pounds lost, and energy expenditure falls [1]. Protein and resistance training can almost completely prevent the muscle loss that comes with dieting or GLP-1 drugs [1]. Consider medication a legitimate option with your doctor, not a failure of character [1].

4. Get children outside early. Time outdoors in roughly the first five years lowers the risk of nearsightedness, a trait many assume is genetic [3].

5. Treat pregnancy and puberty as high-leverage windows. Avoid alcohol in pregnancy, and be cautious about alcohol and cannabis during puberty, given their estrogenic effects [4]. Maternal illness in pregnancy is a known autism risk factor, and infection in pregnancy has measurable effects on offspring [2][3]. Discuss specific exposures, such as herbicides, with a doctor [4].

6. Push for basic nutrition, medical care and schooling for everyone. Linden's point is that people can't reach their genetic potential for height or cognition without them [3].

7. Stop blaming or crediting parents for everything. Twins reared apart show that family upbringing explains almost none of the non-genetic variation in core personality traits, while values, religion and occupations can be passed on [3]. Much of the rest is chance [3].

8. For autism, think in subgroups and ask about biology. Parker's work suggests treatments may help specific children (such as those with low oxytocin or vasopressin) while looking like failures in mixed trials [2]. Ask clinicians about genetically defined subgroups and emerging biomarker research, and rely on evidence rather than disproven triggers like vaccines [2].

9. Swap "nature versus nurture" for Linden's version when you reason about any trait. Ask three questions: what is inherited, what experience (from the womb onward) interacts with it, and how much is simply the randomness of development [3].

Sources

1. Dr Zachary Knight The Science Of Hunger Medications To Combat Obesity · Huberman Lab
2. Dr. Karen Parker: The Causes & Treatments for Autism · Huberman Lab
3. Dr. David Linden: Life, Death & the Neuroscience of Your Unique Experience · Huberman Lab
4. How Hormones Shape Sexual Development · Huberman Lab