The Invisible History of the Human Race

by Christine Kenneally

  • • •

  Millions of people are affected by Mendelian diseases. One birth in every thousand is shaped by a single-gene disorder, of which there are at least ten thousand. Yet as enormous as this number is, Mendelian diseases are considered rare. We have known since Francis Galton, the great innovator, eugenicist, and cousin of Charles Darwin, that not everything we inherit can be explained by a single gene. Now, as the revolution in genomewide association studies enables us to compare the genomes of many people, we know it to be true that most genetic effects are caused by more than one gene. Many traits and common diseases cluster in families, so we should be able to find traces of them in the genome, but so far those traces have been elusive. Some rather large factor is missing from the picture.

  The great paradox of this last crank of the scientific wheel is that while we have finally affirmed the principle that many genes may contribute to one condition, we still don’t know how all those genes do what they do. Now that we have the technology to establish how much influence any one gene has, it appears to be the case that a lot of the time they don’t have that much influence at all.

  Take height as an example. In families height appears to be a strongly inherited trait. In addition, genomewide association studies indicate that it is influenced by at least forty different genes. Yet when scientists tried to understand how these genes underlie the pattern of inheritance, they couldn’t work it out. The height genes were found to only explain 5 percent of the difference in height in the population. Clearly, a lot of something happens between the genome and the person, but as of now we don’t understand precisely what that is. Geneticists call this the problem of “missing heritability.”

  The first explanation, naturally, is other genes. If a condition is the product of many genes, it seems likely that the activity of some genes or pieces of noncoding DNA may affect the actions of others. It may also be the case that some common diseases are shaped by rare mutations that have not yet been tracked down. As advances in medical care keep many more of us alive than hundreds of years ago, rare mutations may be on the increase. The problem of missing heritability must also be due in part to the scope of studies. So far most of the subjects in genomewide studies have been European. As more of the world’s genome is surveyed, the picture will inevitably become more detailed.

  Some common disorders or traits may be explained by idiosyncrasies in the structure of the genome: Segments may be inverted or moved to different spots, and there are many varieties of repeats, like the CAG repeat in the huntingtin gene. New mutations underlie some disorders as well. A small percentage of cases of autism are caused by de novo point mutations, which is to say that while the mutations occur in genes, the condition isn’t inherited. In addition to these dark-matter candidates, there is noncoding DNA. When geneticists discover that differences in DNA correlate with differences in health (for example, people with a certain condition have a T in one spot on the genome rather than an A), they have overwhelmingly found that these significant differences occur not in genes but in the noncoding regions of the genome. Why? They don’t know.

  It is also the case that the environment modulates genes, but which elements of the environment exactly? How well you slept as a child? How well you ate? The absence or presence of certain stressors matters too. Did you grow up in a war zone? Did your family live in poverty? Is anyone in your family an addict? What about your family history of disease and its level of education? Were you exposed to a large number of pollutants? Keep in mind that the way the environment shapes genes isn’t through some vague influence: Everything we hear or see or feel or touch is translated into our tissue by the action of biochemicals of some kind, which should be traceable.

  The lives that our parents and grandparents lived may also affect the way genetic conditions play out in our bodies. One of the central truths of twentieth-century genetics was that the genome is passed on from parents to child unaffected by the parents’ lives. But it has been discovered in the last ten years that there are crucial exceptions to this rule. Epigenetics tells us that events in your grandfather’s life may have tweaked your genes in particular ways. The classic epigenetics study showed that the DNA of certain adults in the Netherlands was irrevocably sculpted by the experience of their grandparents in a 1944 famine. In cases like this a marker that is not itself a gene is inherited and plays out via the genes. More recent studies have shown complex multigenerational effects. In one, mice were exposed to a traumatic event, which was accompanied by a particular odor. The offspring of the mice, and then their offspring, showed a greater reactivity to the odor than mice whose “grandparents” did not experience such conditioning. In 2014 the first ancient epigenome, from a four-thousand-year-old man from Greenland, was published. Shortly after that, drafts of the Neanderthal and Denisovan epigenomes were published. They may open up an entirely new way to compare and contrast our near-relatives and ancestors and to understand the way that they passed down experiences and predispositions. As yet it’s unclear for how many generations these attachments to our genes might be passed down.

  • • •

  Even given our ability to read hundred of thousands of letters in the DNA of tens of thousands of people, it turns out that—at least for the moment—family history is still a better predictor of many health issues. For example, it is the presence of a BRCA mutation plus a family history of breast cancer that most significantly raises a woman’s risk of the disease.

  One of the most practical ideas to emerge out the intricate and complex activities of modern genetics is that we should think about our genes as risk factors. The presence of a BRCA mutation doesn’t indicate for certain that you will develop breast cancer, but it increases your risk. There may be other factors in your life, not just genetic mutations but also family history and personal experiences, that compound the risk further. In the case of Huntington’s disease, the presence of forty or more CAG repeats on the huntingtin gene currently means the risk of death from the disease is overwhelming. But when treatments that slow down the progression of the disease or even prevent its development are developed, the risk assessment will change.

  Consider a less lethal but still important mutation on the F5 gene. People with this mutation will produce factor V Leiden (rather than the normal factor V), a protein that increases the likelihood of excessive blood clotting. If people with this mutation spent a great deal of time on airplanes, they would be especially vulnerable to the blood clots that occur in long-haul-flight syndrome. Armed with this information, they could be scrupulous about getting up and stretching frequently during a flight.

  • • •

  Until a few years ago, the only way for most people to access information about their own genomes was through genetic counseling. The service was offered before or during pregnancy when a genetic disorder was suspected, and for the most part the disorders that were detected were Mendelian, so the information was tremendously consequential. Genetic test results were presented in person by a professional who was trained to educate and assist. Since 2007, however, anyone has been able to send off a cheek swab or spit sample and learn about many of their genetic risk factors.

  23andMe, for example, looked at genetic markers for Parkinson’s disease, multiple sclerosis, and diabetes, as well as many other conditions. It also determines genetic susceptibility to certain drug reactions. I found out that I was likely to be more sensitive than most people to warfarin, a drug that is administered to prevent blood clots in emergency situations, like strokes. Should I need the drug at any point, doctors will ideally minimize the amount so that I don’t bleed too much. I also learned that I was four times more likely than the average person to develop celiac disease. The knowledge made me pay attention to a vague set of gastric symptoms I’d had for more than a year. When I finally went to a doctor and did some testing, I found out that I did not have celiac disease, but I did have a food intolerance that has completely changed
how I eat. The symptoms have now disappeared.

  My husband waited for his 23andMe results with great anticipation, as his mother had died of multiple sclerosis when he was twenty-one. Although many people with multiple sclerosis live compromised but full lives, CB’s mother was physically and mentally devastated to the point where she was unable to recognize her children. Because multiple sclerosis can be hereditary, it was the first condition CB checked when he received his genetic analysis. The report said that he was less likely to develop MS than the average person. That result doesn’t mean he won’t develop MS, but it does suggest that, as far as science can tell right now, he won’t inevitably develop it.

  Despite the obvious utility of such information, there is a serious debate in the medical and genetics communities about whether people should be allowed to access their own health-related genetic data. In 2013 the FDA suspended 23andMe’s health service, and though the company and the regulatory body are now in discussions, it is unclear when the service will be restored. At issue is the accuracy of the information being offered and the belief that genetic information should have special safeguards placed upon it. This is quite clearly true for Mendelian diseases. What if consumers discovered terrifying news about themselves or family members? The risk is not to be taken lightly. But, as we now know, the majority of risk information to be gained from the genome is not Mendelian but relies on many factors.

  Robert Green, a physician-scientist at Brigham and Women’s Hospital and Harvard Medical School, told me that much of the fearful attitude toward genetic information was formed in the cauldron of genetic counseling. “Huntington’s disease has been the paradigm for genetic testing for a long time. It’s really not a very good paradigm, because there’s a certainty to it, and most other genetic, even Mendelian genetic, variations are not fully penetrant.”

  As with almost every aspect of the practical use of genetics, there is currently more opinion than research into many basic questions. Is the discovery of a significant genetic risk by an individual outside the medical establishment as potentially harmful as many people fear? Preliminary studies have shown that what researchers suspect will be devastating news for people may not necessarily be so. Green used to work with the design and implementation of drug-testing clinical trials, so he decided to test the notion of treating information as if it were a drug: Will the information cause benefit or harm or both? The particular information shared in his study was whether subjects had a common variant of the APOE gene, which is connected to a high risk of Alzheimer’s. “We designed the study carefully with many safety features and we initially did the study with the minimum number of people necessary to answer the question. The results of this study and several others which eventually evaluated over one thousand subjects was that among volunteers, disclosure of genetic risk information, even for a disease as frightening and untreatable as Alzheimer’s disease, was quite safe,” he said.

  As far as direct-to-consumer testing results are concerned, there are relatively simple ways to filter information. 23andMe graded its results according to how confident it was of the science, and it shared those confidence levels with its customers. If the genetic risk factors for a serious condition were analyzed, customers had to specifically choose to “unlock” their results, so it was not possible to simply stumble upon them. (23andMe did not test for Huntington’s disease.)

  When I asked Jeff Carroll what he thought of direct-to-consumer genetic testing, his main concern was the unregulated testing of children. With Huntington’s, if there is no question of early onset, it is considered profoundly unethical to test children, as a positive result would severely impact the way that child was treated and how he would feel about himself. Only a child himself, once he reaches the age of eighteen, is legally allowed to initiate such a test.

  • • •

  Even if people aren’t harmed by learning about their personal genetic risks, critics wonder if they will use the information to improve their lives, to support the health of their families, and to reduce the enormous costs of health care in the developed world. Indeed, small studies by Navigenics and other companies have shown that there is not a lot of proactive response when people are informed about their genetic risks. In one study subjects were found to have a higher than normal risk for type 2 diabetes. It is well known that onset of the disease is affected by lifestyle, yet even when the at-risk subjects were given information about their susceptibility, many did not adjust their fat intake or increase exercise or consult medical specialists to minimize their risk.

  The irony here is that this finding supports the argument that most genetic information is not exceptional and should not be treated as such by regulators. It’s well known that any kind of health information—whether it is heart related, weight related, or aging related—often has little impact when it comes to getting people off the couch. Yet no one is suggesting that the medical establishment should stop trying to communicate information about exercise and food to the public, only that it needs find better ways to share it.

  Do people even want to know their results? At a genetics conference in 2012, I listened to a presentation about a series of procedures for subjects who said yes to receiving information about whether they had a gene that was implicated in a certain type of cancer. The researchers found that although everyone initially opted in and made appointments to follow up and receive personal risk information, many didn’t turn up to their first appointment to learn more about the condition, and fewer still turned up for the second, at which they would have received their results. This was interpreted as suggesting that people didn’t want to know what their risk factor was. But what if the declining response was actually an indication that people are busy and that if institutions make it too hard to obtain information, subjects won’t make an effort to pursue it? It may be not the fear of risk but the significant time commitment that’s getting in the way of people’s learning about preventable conditions.

  One researcher surveyed members of a poor community and asked what their response might be to information about their genetic risks. Many were extremely fearful and superstitious, and one man told the researcher that he didn’t want to know his results for fear that simply learning about them would somehow cause the condition to occur. Clearly, education is necessary to help many people understand that genes are almost never by themselves fate. One of the most direct ways to do this is to help people become acquainted with their own genome.

  So far there has been little penetration of genomic data into the medical community. Customers of 23andMe and other direct-to-consumer companies describe taking their reports in to doctors who will often not even glance at the data. “How do clinicians cope with a genomic report? How should a report be designed for them? How will they cope with the targeted findings, and how will they cope with incidental findings?” asked Robert Green. All of these questions lag behind the rapid production of massive amounts of genomic data. “Why should genomic information be different than any other sensitive medical information a physician handles, like medical history, psychiatric history, substance abuse, HIV positivity, sexual orientation? Physicians are privy to all sorts of sensitive information and have responsibility for privacy and translation to their patient,” Green observed. “Unless medicine is foolish and abdicates its responsibilities, it’s really incumbent upon conventional medicine to sensibly integrate this into the practice of medicine.”

  The most important requirement for direct-to-consumer testing is that the information it provides about genetic risk be reliable. In a number of studies agencies of the U.S. government and scientists have compared the results of genetic reports from different testing companies, and in all cases there was divergence in how the companies analyzed the results. Some differences will be inevitable, but for certain disorders the same genome, depending on the company, was reported to be high-risk, low-risk, and no-risk. For critics that finding was damning enough to advise
against any genetic testing.

  Ultimately anyone who participates in personal genetic testing, whether it is historical or health related, should understand that these are thrilling, rich, but early days in this particular science. This moment is to the genomic future as the 1970s and early 1980s were to computers. Back then Steve Wozniak, Steve Jobs, and the unsung heroes of the digital age were tinkering in their garages with proto–personal computers. Now most people use not one or even two but many digital devices on a daily basis. If someone turned off all the computers tomorrow, the world would stop. People who use genetic services need to tolerate a degree of uncertainty, as advancements will be made. Yet there will always be some uncertainty where the genome is concerned.

  • • •

  In 1998 the American Anthropological Association issued a statement about race and physical variation that as of 2014 still stands:

  Historical research has shown that the idea of “race” has always carried more meanings than mere physical differences; indeed, physical variations in the human species have no meaning except the social ones that humans put on them.

  It is remarkable that the idea that physical variations between groups have no meaning beyond the social still has considerable influence. We are all genetic creatures, and our families; small, medium, and large population histories; in-Africa and out-of-Africa experiences; and other massive historical contingencies shape the probabilities that play out in our lives.