Genetics and Medicine
The boy who never stopped growing…
Alpina non-fiction Publishing House presents the book by Australian scientist Edwin Kirk "The Boy who did not stop growing ... and other stories about genes and people" (translated by Maria Eliferova).
How to give birth to a healthy baby? Why does my child have heart problems? Will I develop Huntington's chorea, like my father and grandfather? The topic of genetic diseases is becoming commonplace nowadays, but many people have a vague idea of what medical genetics is. Unlike a therapist who accompanies a patient from the cradle to the grave, a geneticist monitors his life even before its inception and years after its completion. During his professional career, Dr. Kirk had the opportunity to take part in the destinies of thousands of people, and this period coincided with the era of an unprecedented breakthrough in genetics.
The experience of a practicing doctor and a scientist, combined with the gift of a storyteller, allowed the author to write an excellent review of the history of medical genetics and evaluate its prospects, clarifying the most difficult questions for everyone who wants to comprehend the subtleties of this new science with real examples.
We suggest reading a fragment of the book.
Several times a year I give lectures on genetics to medical students. My task is to give them a general idea of the importance of genetics in medicine, refresh their memory of the fundamental principles that they are supposed to have already mastered earlier, and also try to identify for them the immediate prospects of this field of science. If you are studying medicine these days, it is almost certain that one of your future patients will have to undergo genetic tests, and since you will need to write out a referral for testing, receive and analyze its results, it is useful to have an idea of what these results mean. Once this course was taught by Anna Turner — she called it "genetics in an hour", and in subsequent years, when this responsibility passed to me, I found every now and then that I should talk about some new achievements, while the lecture immediately stretched out excessively and had to shorten part of the material.
Nevertheless, I always find time to note that all other areas of medicine should be considered only as subsections of genetics. Almost all human ailments, as well as everything that happens to us that is not directly related to diseases, are based on genetics.
Take, for example, injuries. You probably don't think that the possibility of getting into a car accident or getting in the eye is related to genetic problems. However, imagine, there is an important genetic risk factor, on which the probability of something like this happening to you largely depends. We are talking about the Y chromosome. At any moment, starting from the age of 12 months, when we can usually get up, go and get into trouble, men are more likely to get injured than women.
You may already have some guesses as to why this is happening. The obvious culprit is testosterone, because aggressiveness and impulsivity are qualities that can bring trouble, and testosterone, as you know, increases both.
This factor may well play a role, but probably not everything is so simple: perhaps there are other ways the Y chromosome influences behavior. Of course (with all my desire to put genetics at the forefront of medical science), one should not imagine that the differences in the behavior of men and women are due solely to their innate anatomical and chemical characteristics. Masculinity and behavior of men are influenced not only by genetic, but also by social factors. If you are taught from birth that you should show courage and have a love of adventure, no one will be particularly surprised if you really form these qualities. If you are taught that your place is at home and that quiet, passive activities are best suited for you, of course, you can rebel, but on average it is likely that such upbringing will affect your life choices and keep you from such behaviors that are associated with dangers.
As if to confuse things even more, there are other genetic factors that affect your chances of getting injured, regardless of whether you have a Y chromosome. Not all men have the same risk of injury, and there are many patterns of behavior (and risk) common to men and women. Both men and women differ in the degree of impulsivity; some men (and women) prefer to stay at home and play computer games, while others prefer to jump with a parachute. These differences in representatives of the same sex are also genetically determined to some extent, but it is not so easy to understand them as to blame all the blame on one "defective" chromosome.
In short, the genetics of injury is complex — there is an interaction between genes and the environment. The ratio of both is variable, and there are cases when the environment can completely prevail over heredity. If the most peaceful, adventurous woman in Baghdad turns out to be, she may, due to elementary bad luck, become a victim of a car bomb explosion in the market square. If the most (potentially) reckless and aggressive man turns out to be in an environment where men are brought up as opponents of violence, and other opportunities to suffer are limited (note that I could not give an example of such conditions, but everything is relative), he can live to a very old age without becoming a victim of violence, and die in his sleep.
The results of such an interaction between genes and the environment are easiest to study at the population level, rather than an individual. There is about the same difference between them as between climate and weather: we know that on average men are more likely to be victims of violence, as well as the fact that on average it is hotter in summer than in autumn. However, no one is surprised by a cold summer or warm autumn day; individual days are like individual people. And just as we cannot predict what the weather will be like on a certain day next year, even the most complete knowledge of a child's genotype will not help us to unequivocally predict which personality will grow out of him or even what he will be sick with.
Almost all human ailments at least partially depend on genes. The range extends from the diseases to which this book is mainly devoted (when the breakdown of just one gene or abnormalities at the chromosomal level are enough to cause a hereditary disease), to common diseases like stroke, to which both genes and the environment contribute, and the genetic component is caused not by one gene, but by many. This last option includes many different genetic components, each of which slightly changes the probability of stroke. Some increase its risk, others reduce it. There are probably hundreds or even thousands of similar genetic factors that shift the risks of each of the common diseases; for most people, as a rule, each of them individually plays only a minor role. If a genetic variant increases or reduces your chance of stroke by 20%, this is considered a very significant contribution. In the catalog of genes affecting the likelihood of stroke, 287 different positions were noted, changes in which are likely to have a similar effect, and in most cases this effect is very weak.
The main way to identify the contribution of heredity to common diseases, at least at the moment, is the method of genome—wide association study (GWAS). If you decide to do GWAS, you will need a very large sample of people (from tens to hundreds of thousands) about whom you know something. For example, you know their indicators such as height, blood pressure, whether they had a stroke, etc. You take a DNA sample from each of them and study thousands of sites scattered throughout the genome, which are known to differ in humans. Then these genetic data are compared with known information about people from the sample. The goal is to find a connection between the DNA data and the studied characteristic.
Let's say some people have C (cytosine) in a certain position in their DNA, while others have T (thymine). Having examined the DNA of those who have not had a stroke, we find that 50% of them have a C, and the other 50% have a T. Then we consider a group of those who have had a stroke, and it turns out that among these people 60% have a C in this position, and 40% have a T. That is, in people who have had a stroke, C occurs more often than in the control group. The next stage is to repeat the study anew on the so—called "reproduction cohort" (a new group of people) to prove that the connection established last time is not an error. This is necessary, since at the initial stage of applying the GWAS method, a lot of "sensations" appeared, which later turned out to be statistical noise that had nothing to do with reality. The statistical bar for the second study is slightly lower than for the first, since we are looking for one specific object, rather than considering the entire genome as a whole. This means that so many people are not needed for the second sample, but there is still a lot of work to be done. Suppose that the second time you found something close to the result of the initial study. Congratulations, you have discovered a stroke risk factor!
However, this will hardly help you much. First of all, it may turn out that this genetic variation itself is not directly related to the risk of stroke. She may be an innocent traveling companion who just sits and does not touch anyone somewhere near another mutation - the true culprit. This means that the discovery of such a variation is often just the beginning of a long and tedious search for a real villain. The second problem is that the data obtained, in fact, does not tell us anything about a particular person. If half of the population has a C in a certain place of DNA and this option only slightly increases the risk of stroke, there is no need to worry too much if you find it in yourself. The data may not be applicable to another population: for example, the relationship between the presence of variant C and stroke is reliable only for Europeans. Unfortunately, we have an overabundance of studies on European material (I say "unfortunately", since we desperately lack similar studies on the material of other populations).
There is a fairly high probability that the subject of your searches using GWAS is not in this gene. Most often it happens. Sometimes this happens due to the notorious phenomenon of an innocent fellow traveler — there are changes in a significant gene, and C (instead of T) is located next to this mutation. However, most often, when it is possible to establish a causal relationship between this variant and the disease of interest to us, it turns out that the matter is in the control of genes, and not that the mutation of the gene somehow changes the protein encoded by it. The genome is replete with sequences that play an important role in regulating activity in the cell nucleus. This is usually done using signals in the form of an RNA molecule, very similar to DNA, but still different from it. It activates something that suppresses something else, and that, in turn, changes the activity of the gene so that it produces more or less protein... and all this may have to do with the question you are initially interested in. We are still far from fully understanding this signal network, but, apparently, many of the coincidences identified with the help of GWAS are associated with subtle shifts in equilibrium in a very confusing information web - and this is not something that can be found out in a rush.
Ideally, of course, I would like to identify all the genetic variations that cause diseases in humans, as well as various signs, such as growth. If we could finally understand the genetic factors that determine, for example, whether a given person will have a heart attack, perhaps we would have new ways to prevent it.
Long before the sequencing of the human genome, there were attempts to find out how genes determine various diseases and signs. During the notorious dispute "nature or environment" (nature vs nurture), they tried to measure the role of "nature" in practice. In this field, such a parameter as heritability is widely used — the extent to which the variability of a given trait in a population is due to genes, not the environment. The name "heritability" suggests that we are talking about a direct measurement of the contribution of "nature" to this equation, but this is not quite true: in fact, we are talking about variability within a population. For clarity, imagine that you are studying the hair color of two different groups. One group consists exclusively of Nigerians, the other is a random sample of Brazilians. All Nigerians have black hair, so there is no variability in hair color that can be measured, and therefore heritability will be zero. This does not mean that genes do not play a role in determining the hair color of Nigerians — they even play a very important role. At the same time, Brazilians have all shades — from black to blond. This variability is mostly explained genetically, so heritability will be high.
There are several ways to calculate heritability. A common and relatively simple method is to compare how different identical and fraternal twins are. It is assumed that a pair of twins grows in the same environment, up to the conditions of intrauterine development, and that the impact of the external environment on identical and fraternal twins is the same. Since identical twins have all the genes in common, whereas fraternal twins have on average only half of the common genes, it can be expected (usually it happens) that identical twins resemble each other more than fraternal twins, not only externally, but also in parameters such as height, blood pressure, etc. Quite simple — by mathematical standards — calculations allow us to measure this difference within a group of twins and use it to assess heritability. Heritability indicators, regardless of the calculation method, vary from zero (there is no contribution of genes to the variability of this population) up to one (variability is entirely due to genes), and can also be expressed as a percentage.
Estimates of the heritability of various traits vary in different studies and for different populations. One of the parameters, the data on which agree quite well in a number of studies, is growth: in populations where people eat well, the heritability of growth is 0.8. Most of the variability of growth is due to genetics. Chinese studies have shown a lower heritability of growth — about 0.65. This does not negate the important role of genes, but suggests that the environment in China may play a more significant role than, for example, in the United States. A possible explanation is due to the fact that if your mother was starving while carrying you, and you were also malnourished as a child, then you may not achieve the growth to which you would have grown under normal conditions. When studying people whose childhood fell on hard times, there is a stronger influence of the environment, which reduces the value of genes.
The use of GWAS became possible in the mid-2000s, and this method has been widely used since about 2007. Very soon, GWAS specialists were faced with a discouraging problem: heritability was not detected. By 2010, after considerable efforts, only 5% of the growth variability was explained using the GWAS method, which is very far from the 80% predicted by heritability calculations! Over the past ten years, this gap has gradually decreased, but it has not gone away and has not received a full explanation. The largest study of growth genetics at the moment used data from almost half a million UK residents who voluntarily provided DNA and detailed personal and medical information within the framework of the UK Biobank project. A group led by Steven Xu (this is another Xu, we'll talk about him later) managed to use this storehouse of data to explain 40% of the variability of growth — an impressive breakthrough, but we are still far from accurately predicting the growth of a particular person only by DNA. The prognostic method developed by them made it possible to predict the growth of the majority of people in the group (it did not include those from whom the initial data was collected) with an accuracy of several centimeters. It sounds impressive, but the spread of possible real growth values for each predicted growth value was quite large. Imagine the testimony of a witness in court: "I think that the criminal was 173 cm tall, give or take a few centimeters, but perhaps his height was 158 or 188 cm."
In addition, Xu and his colleagues have convincingly demonstrated that, in relation to the growth of people, an increase in the sample size or the number of genetic markers under consideration does not increase the accuracy of prediction. They used this approach as widely as possible. For example, scientists also examined the level of education, which was measured on a six-point scale (the highest point was "having a degree"), and it turned out that they were able to explain only 9% of the variability of this parameter (however, a larger study has a good chance to explain a large proportion of the variability).
So why are even the most complete studies on a huge amount of material able to explain only part of the desired variability? Two main explanations are put forward in this regard. The first suggests that traditional estimates of heritability are significantly overstated: For example, a recent study based on material from Iceland, where genetic data (obtained by genome-wide sequencing) were used to calculate heritability, revealed much lower values than traditional approaches. For such a trait as height, the new heritability score was not 0.8, but only 0.55. Thus, Xu's group seems to have identified over 70% of what can be detected at all. For Body Mass Index (BMI) the values were 0.65 according to the old method and 0.29 according to the new one; for the level of education — 0.43 and 0.17, respectively: in both cases, the difference is very significant.
Perhaps even more interesting than the version of the erroneous calculation is an alternative explanation for the lack of heritability, according to which there are a huge number of yet undiscovered genetic factors affecting the traits being studied, it's just that the influence of most of them is so insignificant that it is too difficult to measure it even in a very large sample of people. This idea is not new — in 2018 she celebrated her centenary. In 1918, Ronald Fisher, one of the founders of modern statistics, proposed a model of infinitesimal quantities (also known as the "infinitesimal model"), according to which variable traits, such as height, are controlled by an infinitely large number of genes, each of which has an infinitesimal influence on the trait, along, of course, with the influence of the environment. Fischer did not mean that there are actually an infinite number of genes, he offered only a certain approach to the problem.
Portal "Eternal youth" http://vechnayamolodost.ru