TCS Daily

Mankind Version 2.0?

By Charles Murtaugh - May 20, 2003 12:00 AM

A recent University of Michigan survey brings dismaying news: despite abundant press coverage, the public knows so little about genetics that, as blogger Kevin Drum observes, untrained chimps would have done better at answering five simple true-false questions on the subject.

A peek at the actual questions, though, puts one in a more forgiving mood: rather than ignoring media coverage of genetic advances, lay people may have been paying too much attention to the hype. Consider question 3: you mean to tell me that we can't predict heart disease risk from a gene test? Or question 2: there isn't any genetic test for depression? Really? What's all the fuss about, then?

In the debate over whether genetics promises good things or bad, both sides agree on one thing: science is on the verge of remaking mankind. From the editorializing of Leon Kass and Francis Fukuyama on one side, or Gregory Stock and Lee Silver on the other, it's no wonder that the public expects - and dreads - so much from the field. Now environmental writer Bill McKibben has entered the fray with a new book, Enough: Staying Human in an Engineered Age. McKibben paints a picture of molecular genetics about to seize the reins of human evolution:

Last spring an Israeli researcher announced that he had managed to produce a featherless chicken. . . Now, that engineer was not trying to influence his chickens to shed their feathers because they'd be happier and the farmer would be happier and everyone would be happier. He was inserting a gene that created a protein that made good and certain they would not be producing feathers. Just substitute, say, an even temperament for feathers, and you'll know what the human engineers envision.

The above is from an Orion magazine excerpt of Enough. With the possible exception of some token egalitarianism, McKibben's arguments are very much of a piece with those of Kass and Fukuyama, and like theirs it makes good use of the other side's excessive claims:

"With reprogenetics," writes Lee Silver, "parents can gain complete control over their destiny, with the ability to guide and enhance the characteristics of their children, and their children's children as well." [J]ust like the chicken guy, they would be inserting genes that produced proteins that would make their child behave in certain ways throughout his life.

In turn, the proponents of Mankind Version 2.0 hold up McKibben and Kass as repressive Luddites; it's a great help that one of them has the ear of the White House. This certainly makes for lively arguments and book sales (not to mention the advancement of more stealthy agendas), but it is increasingly divorced from scientific reality. A few recent stories from the actual front lines of genetics might serve to show just how implausible the post-human future remains.

First, the featherless chicken. McKibben claims that the researcher "inserted a gene" into the hapless fowl, but in fact it was derived by conventional breeding of two strains, the same sort of thing that could have been done at the turn of the last century. Although the scaleless mutation, causing the lack of feathers, results from alteration of a single gene, that gene has not yet been isolated and remains outside the realm of recombinant DNA.

One day, of course, it may well be isolated: this is what molecular genetics is good for, identifying the molecular basis of very strong genetic effects. When mutation of a single gene causes a dramatic phenotype or disease, such as cystic fibrosis or retinoblastoma, it can be readily mapped and isolated. Genetics is much less useful when the trait in question is hard to objectively define and measure, and when multiple genes make small contributions to the trait.

Consider the following analogy: you are standing in an orchard at night, a photographer's light meter in your hand. In front of you is a tree, giving off a certain amount of light as measured by the meter. That light represents the genetic component of the trait in which you're interested. It may come from a single bright lightbulb hung in the tree; without even looking up, you would be able to zero in on the bulb just by watching the light meter as you train it on different branches. The closer to the light, the higher the numbers on the meter. This is analogous to mapping a single gene trait, such as cystic fibrosis.

On the other hand, that same amount of light could be emitted by dozens of tiny Christmas tree bulbs; at a distance, the meter's reading is the same as with a single bulb, but as you move closer the numbers go haywire: now they are high, as the meter breezes very close to one bulb, but then it drops off with a single twitch of your wrist. Your frustration mounts: you know that there is light coming from the tree, but you can't pin down the sources.

This is the frustration felt by researchers trying to identify genes influencing complex human traits such as personality and intelligence. Consider the group led by Robert Plomin of London's Institute of Psychiatry. In 1998 they published a landmark paper pinpointing a change in a single human gene, IGF2R, as the cause of about four IQ points difference between individuals. The New York Times breathlessly hailed the finding as a "key to high IQ," and it featured prominently in Matt Ridley's excellent book Genome. This November, however, with much less fanfare, the authors withdrew their original report, having found that it was not reproducible with a new study population.

Nor is this an isolated example. As behavioral psychologist Steven Pinker noted in his testimony before Leon Kass's bioethics commission,

Anyone who has kept up with the literature on behavioral genetics has noticed that there's been a widespread failure to find single genes for schizophrenia, autism, obsessive-compulsive disorder, and so on. And those, by the way, are the areas where we're most likely to find a single gene simply because it's easier to disrupt a complex system with a single defective part than it is to install an entire complex ability with a single gene. The failure to find a gene with consistent effect on, say, schizophrenia means that it's even less likely that we will find a gene for something as complex as musical talent or likability.

When the trait is complex, and the genetics are complex, their interaction may well never be unraveled. Nor is this problem confined to human genetics. Consider a recent Nature Neuroscience paper on mouse behavior, by Darlene Francis and colleagues. They were studying two inbred strains of lab mice that exhibit consistent and specific differences in particular behaviors, such as performance in a maze navigation test. The conventional approach, consistent with the genetic determinism prevalent in the human genetics debate, would be to interbreed the strains, and use genetic mapping to fish out and isolate the genes responsible for the apparently heritable differences.

Francis et al. took a very different approach, actually transferring embryos of one strain to the womb of the other and asking which influence predominated, that of the genetic mother's DNA or the birth mother's environment. Remarkably, they found that the differences in adult behavior correlated not with their genes but with the womb in which they underwent fetal development.

The implication for human couples who pay top dollar for "Ivy league eggs" should be obvious. Less obvious, but equally profound, is the implication for human gene mapping. If the mouse behaviorists had followed the conventional cross-and-map approach, they likely would never have found responsible genes. Instead, their work would have been fraught with one false lead after another, just as have afflicted human behavioral geneticists.

Is it theoretically impossible that we will eventually use genetics to "control" the psychological characteristics of our offspring? No, but if one looks at the rather sorry state of human behavioral genetics and then considers the daunting technical hurdles to changing genes, even if they were identified, then the prospect of a post-human future becomes much more remote. A thousand monkeys might eventually type out the works of Shakespeare (or perhaps not), but we shouldn't base primate conservation policies today on that possibility. Similarly, comparing the overwhelming claims of horror and promise in human genetics to the underwhelming reality, one might decide to base policy decisions on the science of today rather than the science fiction of a million tomorrows hence.

TCS Daily Archives