An urgent plea for a broader understanding and awareness of the unconsidered dangers of new genetic technologies. Since 2010 it has been possible to determine a person’s genetic makeup in a matter of days at an accessible cost for many millions of people. Along with this technological breakthrough there has emerged a movement to use this information to help prospective parents “eliminate preventable genetic disease.” As the prospect of systematically excluding the appearance of unwanted mutations in our children comes within reach, David B. Goldstein examines the possible consequences from these types of choices.
Excerpts from the author’s latest book, The End of Genetics, published by Yale University Press
by David B. Goldstein
The first decade of the twenty-first century marked a turning point in the relationship between society and the human genome. For the first time it became possible to determine the genetic makeup of any person in a matter of days and at a cost already within range for many millions of people. Even before this genomic watershed was reached, a movement had emerged to provide genetic information directly to consumers.1 In some cases the offerings to consider included help to make “more perfect babies.”2 The obvious question upon reading such a claim is this: What would qualify as a more perfect baby? Until recently, this question was of primarily academic interest. Ethicists could debate the benefits and the risks, but there was no realistic prospect of systematically prohibiting the appearance of unwanted mutations in our children. For better or worse, the genomic engineering of future generations has suddenly become a very real prospect and is therefore something, I believe, that society must urgently consider. There are two key questions about such reproductive genomic engineering that need to be considered and understood: What do we want to change in the genomes of our children, and what are we likely to be able to change?
In terms of what we would like to change, in my experience, most people are comfortable with the idea of ensuring that children do not carry mutations that cause devastating childhood diseases. Indeed, there are already ongoing efforts throughout the world to reduce the transmission of severe childhood genetic diseases such as Tay-Sachs disease. Many couples already opt for a procedure called carrier screening, which seeks to identify genes for which both mom and dad carry mutations that, when combined, would result in a severe childhood genetic disease. And many fertility centers offer the option of testing embryos for mutations that would cause such diseases in order to select for transfer those that do not carry two such mutations, among other genetic testing options.
But how far should these efforts go? Is it also appropriate to test for mutations that predispose one to, but do not deterministically cause, later-onset conditions such as Alzheimer’s disease? And what about non-disease traits, such as height or eye color or, indeed, traits such as sexual orientation, where the genetics is often only probabilistic but where those probabilities can in principle be tested for and parental preferences acted upon? Even if we restrict attention to disease, we face the question, What is or should be classified as a disease? These questions become particularly acute when we recognize that different people will have very different ideas about what they would like to see in the genomes of their children. And views about what is and is not a disease have changed in important ways over time. As one stark illustration, a certain minority of the deaf community has a preference for deaf children. What if a couple wants to use genomic technologies to ensure that their children would also carry the same deafness mutations that they themselves carry? In the past, professional geneticists could draw a certain comfort from technological constraints. As our ability to identify the mutations that make us different from one another grows along with our ability to not only select but, as we shall see, edit what is present in the genomes of our children, what we should and should not do will emerge as one of the defining questions for societies and individuals.
The second part of the question about what we might change in the genomes of our children relates to what kinds of traits we are able to influence and how much we can influence them. When the personal genomics movement was first gaining traction, it was heavily focused on common genetic variants, mainly for technical reasons related to how genetic studies at the time were performed. These common variants are conventionally considered sites in the genome where there is a common form and a minor form, and the minor form is observed about 5 percent or more of the time. This represents a tiny fraction of the variable sites in the human genome. As you will learn in some detail through the course of this book, there are around three billion different positions in the human genome, and we know that most of these three billion sites vary in one or more humans alive today. As you will also learn in the chapters that follow, while we know that most of these sites will vary somewhere in the human population, we still do not know the consequences of the vast majority of that genetic variation.
In the early days of personal genomics, there were two clear drivers of consumer interest: genetic ancestry and the results of “genome-wide association studies,” or GWAS. Genetic ancestry is just what it sounds like—your genetics telling you something about where your ancestors are from in a geographic sense and, sometimes, something about relatives you may not have known you have. And GWAS is a fancy, and somewhat overstated way of describing a type of genetic experiment that allows the assessment of whether any of the common variants in the human genome influence a particular disease or trait. A typical GWAS might, for example, compare a million common variants between patients with schizophrenia and those without schizophrenia to see if any of them are associated with an increased risk of disease. As of this writing, many thousands of different common variants have been associated with hundreds of different diseases and traits, including, by now, almost all the common diseases and many other traits such as height and weight, skin and hair color, and even complex behavioral traits such as educational achievement, IQ, and sexual orientation. All of this can be reviewed using a catalog of GWAS findings.
In the case of ancestry testing, consumers could learn something, albeit rather coarse, about the geographic origins of their Y chromosomes or their mitochondrial genomes, reflecting, respectively, their paternal or maternal ancestry. Or they might get a composite picture of a kind of average of the various geographic ancestries represented in their entire genomes, learning, for example that their ancestry overall appears to be from some part of Europe, Asia, Africa, or the New World. Or they could learn that they are related, to some degree, to someone else who has also been tested. In the case of comparing their genomes against the results of GWAS studies of diseases and other traits, consumers might learn that they have a marginally greater or lesser risk of type 2 diabetes than the population average or that they carry stronger risk factors for certain autoimmune or neurodegenerative diseases. In most cases, however, there has been no reputable advice that could be offered as a function of the genetic information gleaned from such “genomic profiling.” For example, one of the very strongest effects for any common variant in the human genome is due to variation at the ApoE gene. Those who carry the risk forms of this gene have a greatly increased likelihood of developing late-onset Alzheimer’s disease, but there is currently nothing that carriers can do to meaningfully alter their risks of developing the disease.
This lack of clinical impact led me, in a New York Times interview in 2008, to sum up the entire enterprise as “recreational genomics.” Referring to the field as recreational was not only intended to convey that what was being discussed in the guise of “personal genomics” was of little to no clinical value but also reflected a degree of exasperation that I (and many other geneticists) feel about how many personal genomics companies implicitly or explicitly oversell what they offer and, in consequence, badly mislead the public about the nature of human genetics. Evidence of how personal genomics companies oversell and misinform is not hard to find. Sometimes it is subtle; sometimes it is egregious.
Alistair Moffat is a journalist and former rector at St. Andrews College in Scotland. He is also the former chief executive officer of Britain’s DNA, a now-defunct for-profit personal genomics company that once offered consumers a variety of DNA tests. In the summer of 2012, Mr. Moffat described some of what Britain’s DNA had discovered. He first claimed that a volcano seventy thousand years ago “blew itself to smithereens” and destroyed all the human genetic lineages except for those of two individuals, dubbed Adam and Eve. He went on to claim that Britain’s DNA discovered a remarkable individual who has Eve’s DNA, that they had found a genetic marker from “Queen Sheba,” that 33 percent of the men of Britain carry the “founding lineages of Britain,” and that 97 percent of men with the surname Cohen share a genetic marker. The earlier description of what some personal genomics companies offered as “recreational genomics” can hardly cover excesses like this. An English geneticist, Mark Thomas, who was involved with me in some of the work that provides the grain of science behind these absurdities, more accurately described them as genetic astrology. Astoundingly, Mr. Moffat’s initial reaction to being challenged about the accuracy of these claims was to threaten legal action against Dr. Thomas and one of his colleagues, Dr. David Balding, a highly respected and talented mathematical geneticist. Mr. Moffat’s science-free musings would almost be amusing if their potential consequences were not so serious.
The egregious misrepresentation of genetics for what seems to be commercial gain reflects an increasingly complicated relationship between professional geneticists and a public eager to know what the vaunted genomics research might mean for them in terms of their personal histories and the diseases they might face. When the personal genomics craze got started, many professional geneticists felt there was little that was really useful for anyone to know from having their genomes profiled in the ways possible during the dawn of personal genomics. It was very awkward, however, for geneticists to insist that people have no reason to learn anything about their genetic makeups if they are interested in doing so.
David B. Goldstein is John E. Borne Professor of Genetics and Development and director of the Institute for Genomic Medicine at Columbia University Medical Center. He is the author Jacob’s Legacy: A Genetic View of Jewish History.