Innate, page 4
Finally, it is important to emphasize that heritability is not the same as heredity or inheritance, or at least not always. For animal breeders, heredity is the important aspect—how strongly offspring resemble their parents. But heritability actually measures all genetic influences on a trait, not all of which are actually inherited. First, many traits are caused by multiple genetic factors acting together—the particular combinations of genetic factors may be crucial in determining the phenotypic outcome in each individual. Because each of our genomes represents a new combination of those genetic variants, these will be different from either of our parents. Second, we each also have new mutations in our genomes that arose in the generation of the sperm and egg cells from which we were formed. These also contribute to our individual traits but obviously do not contribute to parent-offspring similarity. Down syndrome provides a stark example of this; it is a condition that is rarely inherited from a parent—it most often derives from a new event in the egg or sperm that leads to an extra chromosome 21 being included—but it nevertheless has a completely genetic mechanism in the individual. Both these factors—the influence of new mutations and the importance of unique combinations of genetic variants—make large contributions in twin studies as MZ twins share all new mutations and also the exact same combinations of all genes.
NONGENETIC EFFECTS
I have been emphasizing the heritability of psychological and brain traits in humans, but twin and adoption studies also highlight nongenetic contributions to overall variance. These effects are often assumed to be “environmental” in origin, but we will see that that is not necessarily the case. The same comparisons of MZ and DZ twin pairs or biological versus adoptive siblings that are used to calculate heritability can also be used to estimate the variance explained by different family environments.
Consistently, and surprisingly, this turns out to be very low (usually not more than 10%–15%) and is often found to be zero. Generally speaking, adoptive siblings do not resemble each other for psychological traits any more than two strangers in the street. This is despite being raised in the same household, living in the same community, typically attending the same schools, and so forth. And for many traits, MZ twins who are reared apart are almost as similar to each other as those who have been reared together—sharing a family environment does not make them appreciably more similar.
This result has caused some consternation and even disbelief over the years since it was first highlighted by, for example, Judith Rich Harris and Steven Pinker. However, it is actually far less surprising if we consider the kinds of traits we are talking about. They are the very ones that, by definition, reflect some stable differences between people, some underlying dispositions that influence patterns of behavior over time. Any parent with more than one child will likely have noticed differences between them that cannot be traced to differences in parenting—in fact, these are an endless topic of conversation between parents. Why is one child studious and attentive while the other has his or her head in the clouds? Why is one cautious while the other is on a first-name basis with staff at the emergency room? Why is one so shy and quiet that you worry he or she will never have any friends while the other would happily stand talking to a post? Children have different temperaments, different talents, and different interests that simply seem to emerge of their own accord and to be largely resistant to any efforts to change them.
Academics love to find things that are counterintuitive—that conflict with our everyday experience and show how wrong we can be about the way our minds work. This is not one of those cases. The results from twin studies do not actually conflict with our intuitions and our common experience at all. These studies are about precisely those kinds of traits that we encounter as parents as largely innate or intrinsic to the child. We should not be surprised if the results fit with this experience.
Does this mean parenting doesn’t matter? Of course not. It doesn’t even mean that parenting doesn’t affect our offspring’s behavior—of course it does. Love, encouragement, support, discipline, expectations: all have hugely important impacts on children’s lives. They shape the characteristic adaptations we all have to the situations in our lives, to our expectations of ourselves and the choices we make. It just means that parenting doesn’t significantly affect their underlying behavioral traits or predispositions. But those traits are only part of what influences people’s actual behavior.
These findings suggest that many reported correlations between parental behavior and offspring traits do not reflect a direct causal link, as often inferred, but instead reflect the effects of shared genes. If, for example, we find that overprotective parents have anxious children, this could be because overprotective parenting causes children to be anxious. But the evidence described above is not consistent with such an interpretation, as it should affect MZ and DZ twins or adoptive and biological offspring equally. Instead, the general findings suggest that parental overprotectiveness and child anxiety are more likely both manifestations of the same genetic effects, acting in both the parents and the offspring. Similarly, growing up in a household with more books in it is correlated with higher IQ—does this mean reading raises your IQ? Well, I’m all for reading, but this correlation more likely reflects the fact that parents with higher IQ tend to have more books in the house and also tend to have children with higher IQ. In general, these kinds of sociological correlations are thus hopelessly confounded by possible (indeed, likely) genetic effects.
It should be stressed, however, that most twin and adoption studies sample a relatively small range of potential family environments. Many studies have shown that serious neglect or abuse can have long-lasting psychological consequences. Fortunately, such situations are rare, at least rare enough that they do not make much of a contribution to overall variance in psychological traits across the population and thus do not show up in the shared family environment component of variance. Again, these studies only measure the factors that actually do make a contribution to variance in a population—not all the ones that could make a difference, if they occurred.
If genetic effects account for 40%–60% of the phenotypic variance and family environments account for only 0%–10%, that clearly leaves a good chunk of the variance unexplained. Something else is making people different from each other, even MZ twins who grow up in the same family. That factor is referred to as the “nonshared environment,” but we will see in later chapters that much of this may be caused not by any factors outside the organism, but by inherent variation in the processes of development themselves.
In the next chapter, however, we will concentrate on the genetic effects. We will look at what genes actually are, where genetic variation comes from, and how it affects the kinds of traits we are interested in.
CHAPTER 3
THE DIFFERENCES THAT MAKE A DIFFERENCE
When we say that genes influence behavior, what we really mean is that genetic differences contribute to differences in behavioral traits (which in turn influence patterns of behavior over time). So, what are these genetic “differences”? To answer that, we need to start with a more basic question: What are genes?
You might think there is a simple answer to that question, but there isn’t. Defining what a gene is has in fact been a source of enormous confusion both within science and for the general public. The reason is that the term actually refers to two very different things. The original concept, famously devised by Gregor Mendel in the 1850s in studying various traits in peas, was of some physical thing that gets passed on from parent to offspring, and that determines the trait in question. From the patterns of inheritance he inferred that there must be distinct genes for whether peas had smooth or wrinkled shells, whether they were green or yellow, whether the flowers were white or purple, whether the plants were tall or short. He also was able to deduce that each plant inherited two copies of each gene—one from the mother and one from the father. Importantly, Mendel realized that each of these traits was controlled by a discrete inherited unit—different ones for different traits. The term “gene” was introduced later to refer to these units of heredity.
While Mendel knew that these units must have some physical substrate—genes must be a physical thing—he didn’t know what they were made of. It was not until the 1940s that scientists figured out that the genetic material was DNA—deoxyribonucleic acid, a major chemical constituent of the chromosomes (literally, colored bodies) that were visible down the microscope in the nucleus of cells.
This fact is so well known now that it’s hard to think of a time when it wasn’t, but actually DNA was not even a front-runner in the betting for what substance carried the genetic information. It was deemed too simple, as it is composed of only four different chemical subunits, or bases, arranged in a long sequence along each chromosome. The preferred candidate was proteins, also present in chromosomes and throughout cells—these are much more complicated than DNA, as they are composed of 20 different amino acids strung together in long chains, which then fold back on themselves to form complicated three-dimensional shapes. While DNA just kind of sits there, proteins are properly impressive—they do all sorts of things inside cells, acting like tiny molecular machines or robots, carrying out tens of thousands of different functions.
Proteins thus seemed a much likelier candidate than DNA to be the genetic material. But a seminal experiment looking at how one type of bacterium could be transformed from a nonvirulent to a virulent (disease-causing) form clearly showed that it was DNA and not proteins that carried this genetic information. (It turns out the proteins associated with chromosomes are involved in packaging the DNA inside cells and in regulating which genes are expressed, but do not themselves carry the genetic information.) From our vantage point, in the digital age, this now seems unsurprising. The simplicity of DNA that led many to dismiss its information-carrying capacity can now be seen as ideal if the information is carried in the sequence of the bases that make it up, just as it is carried in a sequence of 1s and 0s in a computer. Moreover, the fact that it is chemically very inert—it just doesn’t do much—is exactly what you want in order to safely and stably encode information over long periods, not just over the lifetime of an organism, but also over many generations spanning millions of years.
In fact, these properties were predicted on theoretical grounds by the physicist Erwin Schrödinger in a famous series of lectures on “What Is Life?” delivered in 1943 at Trinity College Dublin. He realized that what set living things apart from nonliving ones was that living things are organized. Both living and nonliving things are made of the same kinds of stuff—of atoms—it’s just that in living things these atoms are organized into molecules and complexes of molecules and cells and organs. Keeping things organized is hard work, as the general trend in the universe is for things to get messier, if left to their own devices. It requires energy to keep things organized, which all living things must take in, in some form or another, but it also clearly requires information. An organism must contain within it the information for how all those molecules and cells should be organized. And it must be able to replicate that information and pass it on to its offspring. Schrödinger realized that what he called an “aperiodic crystal” would be a perfect medium to store such information—that is, the material should be stable, like a crystal, and should contain within its structure a code, written in the nonrandom, nonrepeating sequence of different subunits.
THE STUFF THAT GENES ARE MADE OF
DNA fits that bill perfectly. The most obvious and direct thing encoded in DNA is, a little ironically, proteins. The recipes for all those impressive micromachines whizzing around in our cells are written in our DNA. And this brings us to the second definition of a gene—one derived from molecular biology, rather than the study of heredity. Here, a gene is a stretch of DNA that codes for a specific protein. Each chromosome in the cell is a single continuous molecule of DNA, like a long string, made of a series of the four different chemical subunits joined together. These subunits are called adenine, thymine, cytosine, and guanine, but are usually referred to as the “letters” of the DNA code: A, T, C, and G, respectively. Each of these molecules has a polarity to it—they have two ends where they can be chemically joined with the other bases—actually rather like the way we join letters together to make words.
The chromosome is made of two of these strands of DNA wound around each other in the iconic double helix. The information on each strand is complementary to the other due to the way that the chemical bases interact with each other: an A on one strand will be matched by a T on the other, while a C on one strand will pair with a G on the other. This gives an obvious mechanism for copying DNA—the double helix can be unwound and the two strands pulled apart, with each one then acting as a template for construction of another version of the other one, yielding two copies of the double helical molecule.
If you start at one end of a chromosome and scan along it (on one strand), you will soon come to a bit of the DNA that is special, because the sequence of bases here encodes a protein. That is, the sequence of letters is a code that tells the cell which amino acids to string together, in what order, to make protein A or protein B, and so on. It took a while to work out, but we now know that each successive three-letter stretch of the DNA sequence corresponds to a different amino acid. There are also three-letter codes that tell the cell where the code for a particular protein starts and where it ends. So, if you keep scanning along the DNA, you will also come to the end of the section that codes for whatever that protein is. Figure 3.1 illustrates the structure of a gene.
Figure 3.1 The physical structure of genes. Spread out along each chromosome are genes—stretches of DNA that code for proteins. The sequence of DNA bases—A, C, G, and T—codes for a sequence of amino acids in the corresponding protein. DNA sequence in regulatory regions controls protein expression.
From a molecular biological point of view—the perspective that aims to understand how cells work rather than how traits are inherited—that stretch of DNA is a “gene.” We have about 20,000 different genes spaced out along our 23 chromosomes, which collectively make up the human genome. They code for proteins like collagen, hemoglobin, insulin, metabolic enzymes, antibodies, ion channels, neurotransmitter receptors—all the things that cells need to do their various jobs.
TURNING GENES ON AND OFF
Now, things are about to get more complicated. When I said that a gene encodes a protein, that is true, but the gene itself doesn’t make the protein. As I mentioned above, DNA is an incredibly inert molecule—it just stores the information. In order for that information to be acted upon, or expressed, it must be read out by the cell and decoded. The machinery that does that is itself composed of other proteins in the cell. (If you’re starting to see a chicken and egg problem, you’re right.)
These other proteins include, first of all, an enzyme that makes a direct copy of the stretch of DNA that codes for a protein. This process is called transcription because the code is essentially the same, though the physical substrate carrying this copy is not DNA, but its cousin molecule, RNA (ribonucleic acid). This RNA copy, called a message, is then transported out of the nucleus of the cell—the information storage compartment—to the cytoplasm of the cell, which is where proteins are made. The RNA message is, like a tape, gradually passed through a complicated molecular machine called the ribosome (made of proteins and other types of RNA molecules) and at each successive three-letter code, the appropriate amino acid is added to the growing string that will form the new protein. This process is called translation because it takes the information in the language of nucleic acids and translates it into the language of amino acids. When the end of the message is reached, the protein is released, folds up into its predestined shape, and flits off to do its job, wherever in the cell it is needed.
But here’s the rub—different cells need different proteins. Blood cells make hemoglobin but other cells don’t. Immune cells make antibodies. Pancreas cells make insulin. Each different type of cell in the body—and there are many thousands of different types—expresses a different subset of the 20,000 proteins encoded in the genome. In fact, that profile of gene expression is precisely what makes muscle cells different from nerve cells or skin cells or blood cells.
So the DNA has to encode much more than just the recipes for each protein or active RNA—it also has to encode the instructions for when to make them, where to make them, how much of each one to make in any particular cell. This information is encoded in the sequence of DNA that flanks the part that encodes the protein itself. It is interpreted by other proteins in the cell, which seek out and bind to short stretches of DNA, promoting or inhibiting production of the RNA message from that gene. Each cell type makes a distinct set of these regulatory proteins that control and coordinate the expression of all the other genes in the cell. I should add that, in addition to protein-coding genes, there are several thousand other genes that encode RNA molecules that are not merely messengers, but that themselves have some active function in cells.
