Continuing watching and writing a summary for Robert Sapolsky’s course “Human Behavioral Biology”. Lecture 6.
Lecture 6: Behavioral Genetics I
Let’s see how the methodology for studying genetic component of behavior changed with time.
When the field started, people assumed that
if a trait is universal in a species, it must be genetic.
This is very limited approach, which works for flies, but does not for more complex organisms.
The next step was to consider families. The closer related two individuals are, the more genes they share. If you have a behavioral trait that becomes more common the more closely related two individuals are, it must have genetic component. However, the more related two individuals are, the more they share their environment as well. Thus, with this approach one cannot distinguish genetic component of behavior from environmental influence. This was a classic approach 70-80 years ago.
Then let’s control for environment. Examine relatives in the same environment, but having different genes. For example, one can consider identical (monozygotic) twins (sharing 100% genes) vs fraternal (dizygotic) twins (sharing 50% genes). If they all live in the same environment and fraternal twins differ in some way, while identical twins don’t - this must have genetic basis. Problems:
- Sometimes fraternal twins have different sex -> let’s restrict to same sex fraternal twins.
- These twins are still treated differently as pairs. As a result, there is more environment differentiation for fraternal twins.
- Identical twins can have one or two placentas (based on when the split happened - before or after 5 days since fertilization). In one placenta case, they have same blood flow. With two placentas - there can be differences in blood flow. As a result, in one placenta case they have much more similar environment. For example, they are more likely to have similar IQ when having one placenta (when compared to two placenta case).
What if we have everything exactly the same except sex? Does not work either. 1 hour old boys have more arm movement than girls. However, moms already (from the first time they hold the child) treat them differently (holding for different duration of time and in different proximity to their face).
SAT scores and gender study
One loud example of this is a study about 13 year old pupils’ SAT scores in math. In particular, how they depend on sex. The results were that more boys had higher scores. The paper was published in Nature and concluded that since they had the same lessons (there was no school curriculum differentiation yet at that age), this must be a biological difference. Press quickly called this “the math gene” and it was claimed that boys tended to have it more often. However, there were differences in their environment. Other studies has shown that even in the first grade a boy is more likely to be called when raising his hand for a math problem. Boys are also more likely to be praised for correct answers in elementary school.
Look at individuals raised in the same environment, but with different genes (adoption studies). Consider someone who was adopted as a child. Who are they more likely to share a trait with - biological parents or adoptive? In animals this can be done by cross-fostering, i.e. switching newborns (or only 50% of them) across mothers.
Biological component of schizophrenia
Seymour Kety tried to test whether schizophrenia has biologic component (The significance of genetic factors in the etiology of schizophrenia: Results from the national study of adoptees in Denmark, 1987). The study considered adopted individuals who had schizophrenia. Are they more likely to share this trait with their biological or adoptive parents? Entire database of adoptees in Denmark was researched (Scandinavian countries record a lot of data). In general population probability to develop schizophrenia is 1%.
Probability to develop schizophrenia based on whether biological/adoptive parents have schizophrenia. Numbers are from the course. The original source: Seymour Kety, “The significance of genetic factors in the etiology of schizophrenia: Results from the national study of adoptees in Denmark”, 1987
Yes, having biological parent with schizophrenia increases probability, having both (biological and adoptive) - increases it even further (non-additive synergisms). This was the first study to show heritable basis for a psychiatric disorder. This became a gold standard for behavioral genetics studies. People started applying this approach to depression, alcoholism, criminal behavior. They observed similar results. Problems:
- the adoption does not happen immediately after the birth.
- prenatal environment shared with mom (this can be measured by difference of probability to share a trait with mom and dad).
- paternity uncertainty - sometimes a person claiming to be a father is not a father.
- adoptive family placements are not random (in some countries like US kids are matched across a lot of domains before adoption).
Identical twins separated at birth
The best approach so far is to consider identical twins separated at birth (they can be adopted in different households). This uncovers 50% heritability of IQ, position on introvert-extrovert scale, degree of aggression.
Problem: non-random placements. To mitigate, compare similarities between monozygotic and dizygotic twins (they both have non-random placement), but this makes the sample size small.
Looking at non-learned traits
Can we consider traits which happen without learning? E.g. smiling (even blind babies smile), beginning to babble at the same age (even for deaf babies). Random observation - even fetuses smile during third trimester. Prenatal effects can affect such traits too.
Environment begins earlier than birth
Let’s consider some ways which can affect the child before birth.
Sharing blood with mom
- Rats have dozen of fetuses which have more similar blood environment with adjacent sibling fetuses. As a result, the more male siblings a rat fetus has adjacent to them, the later it will reach puberty (local endocrine effects).
- Children born to younger or older mothers reach puberty later (due to lower estrogen levels in mothers).
- Stress in mother propagates to fetus through stress hormones. In rats this causes smaller brain in adulthood,
thinner cortex, less learning abilities, more prone towards anxiety, stronger cognitive decline in old age.
This also decreases brain region responsible for dealing with stress hormones. As a result such offsprings
have more stress hormones themselves. Their offsprings will be affected even more by this. This environmental effect
can be observed even in grand children (“the grandmother effect”). I.e. this is inheriting of a non-genetic trait
- non-Mendelian inheritance.
Dutch Hunger Winter - almost immediate decline in available food due to war in winter 1944.
Third trimester fetuses learn how many calories are available in the world during a specific time period. They “study” nutritional profile. Then based on the results they can “choose” how their body should function through metabolic programming.
Individuals who were third trimester fetuses during Dutch Hunger Winter (i.e. built their nutritional profile during the famine) were trying to store as much nutrients as possible and their kidneys tried to store more salt as well. When normal diet returned (also almost immediate change) - affected individuals showed 19x increase in incidents of obesity, diabetes, hypertension, metabolic syndrome. This is a very dramatic difference.
This did not happen to newborns or first trimester fetuses. This is now studied as part of “fetal origins of adult diseases”. E.g. stress hormones during fetal life cause higher anxiety in adulthood.
People in Stalingrad didn’t get this effect, because there the food decrease was steady.
What happens when people affected get pregnant themselves? Their bodies take too much calories from blood, and as a result fetus is born with mild version of “Dutch Hunger Winter” phenomenon (again non-Mendelian inheritance).
Females (including in humans) can consume different amounts of estrogen (E.g. vitoestrogen from plants). When consumed more - fetus has consistent increase in estrogen dependent breast cancer in adulthood.
- In rats one can inject sucrose water with particular taste (fetuses drink amniotic fluid). After birth this taste will be preferred when given a choice.
- Mom’s voice in humans. Fetuses hear a lot, but especially moms voice. When reading “Cat in hat” a lot out loud vs random collection of sentences (controlled for length and rhytmicity), newborns prefer “Cat in hat” (they make sucking motion when they like something). When fathers read, this does not work anymore.
Conclusion. Environment does not being at birth.
Remember strains of rats with different levels of anxiety mentioned in previous lectures? There was “an adoption” study, where fetuses were transfered during pregnancy. As a result, it was not a genetic trait, but a prenatal effect.
Genetic influence from mother
If you see more shared traits with mother than with father, these can be attributed to prenatal effects. However, one gets more genetic influence from mother, e.g.
- when father is not real father
- mitochondria has its own DNA. Sperm does not carry mitochondria, but only DNA. As a result, one inherits all their mitochondria DNA from mother.
- imprinted genes - work differently based on whom they come from.
- sperm does not have transcription factors. All transcription factors in a fertilized egg are from mother. Imagine there is a transcription factor A and it is needed to produce more of A and to activate gene X. If due to an environmental change, A stops being produced, there will be no A in the egg and this will disable X for offsprings forever (i.e. genes are inherited, but won’t be expressed due to acquired trait in mother - lamarckian traits).
Indirect genetic effects
There can also be indirect genetic effects:
- Consider 50% heritability of your position on introvert-extrovert scale. This conclusion held through all the concerns above. However, height and appearance are highly inheritable. Taller people are treated better and considered more attractive. People treated more positively during childhood become more extroverted. Most of introvert-extrovert heritability is mediated by heritability of physical traits.
- Heritability of rank in turkeys - actually it is due to the color of their feathers, which is inherited.
- Chicks are able to peck for grabs very early. However, newborn chicks have heritable tendency to find their toes very interesting, as a result they peck at their toes and learn to do proper pecking.
- 70% heritability of political party affiliation in US. Mediating variable - how people feel about ambiguity. Conservatives are more ambiguity averse.
- Rats and mouse strains with various levels of aggression. Mediation: aggressive strains have lower threshold for pain sensitivity.
DNA cover can block transcription factors from reaching DNA permanently. This happened in Dutch Hunger Winter people. Another example: pet newborn rat for 3 min/day for 2 weeks after it is born -> larger brain, better learning. Pet it for 1.5 hours/day instead -> smaller brain, shorter life expectancy. It is interesting what counts as stimulation (3 min) vs stress (1.5 hours). However, this actually changes mother behavior. 3 min petting causes mom to get excited and to lick and groom the kid more. 1.5 hours causes ignorance for long period of time. Petting itself does not matter, but only mother reaction. Licking and grooming by mother causes larger brain and all other possible outcomes mentioned above. This can have multi-generational effect - licked and groomed daughters will lick and groom their kids more ( again non-Mendelian inheritance). This changes access of transcription factors, which are related to activating genes for making receptors for stress hormones. It is also reversible. Cross-fostering after non-grooming mom improves kids state.