NOVA Online (see text links below)
Cracking the Code of Life

Collins Dr. Francis Collins
Meet the Decoders
Dr. Francis Collins

Krulwich: There's an analogy I was given that Mozart would sometimes compose music [such that] you would put the sheet of music in front of you and you read from the top to the bottom. You'd play a beautiful tune. Then you could take the music and simply turn it upside down and read it in what is in effect a mirror image and play another beautiful tune.

Collins: That sounds apocryphal to me, but it's a great story.

Krulwich: Is that the sort of thing? That is, depending upon how you read the genes you can get different end products?

Collins: Maybe think of it more as if you are reading a mystery novel, and there are several chapters, and you could pull some out or stick others in. People actually write novels of this sort. You can make the ending different if you get sick of that particular one. There's an alternative way to substitute a few pages here and there.

That's sort of what the gene is doing. You read it one way and you get this ending; you change around a few chapters, a different guy did it. It's the same kind of parallel. You have enough information there to code for several possible outcomes.

Krulwich: What does folding have to do with it?

Collins: Well, genes are effectively one-dimensional. If you write down the sequence of A, C, G and T, that's kind of what you need to know about that gene. But proteins are three-dimensional. They have to be because we are three-dimensional, and we're made of those proteins. Otherwise we'd all sort of be linear, unimaginably weird creatures.

So the DNA information, in the process of being translated into proteins, you end up making a string of amino acids. That's what a protein is. But they're not going to be happy lying out flat in a long, linear array. They have needs and reasons to want to be snuggled up against each other in a particular way. And actually a particular amino acid sequence will almost always fold in a precise way. Every time you make that one it will fold up in a certain way, with some help from the cellular machinery.

Krulwich: Shall I think origami-like, an interesting folding?

Collins: It's very elegant, very complicated. And we still do not have from first principles the ability to precisely predict how that's going to work. But obviously it does work by a combination of chemistry and a little nudging maybe from the cell saying, "Oh, don't twist that way, twist that way, and then you'll get it right." But it does work. You end up with proteins that virtually always come together in this three-dimensional shape, without which they couldn't do their work.

Krulwich: Because they want to stick here, land here, point here or what -- shove there or something.

Collins: And if you have a long sequence of amino acids, it may be that this one over here has to be next to this one over here or the critical thing that protein has to do can't happen. So they have to hold up just right so that they actually end up next to each other in the three-dimensional structure, even though they were far apart in the beginning.

Krulwich: So to identify these different proteins you not only have to know what they're made of but you have to fold them properly?

Collins: Yes.

Krulwich: And to solve a mistake, to make a bad protein good, sometimes it would just involve re-folding it?

Collins: Well, cystic fibrosis is a very well-known example where the problem seems not so much to be that there is a missing amino acid (although there is), but that the protein as a consequence of that doesn't fold right. So if you could get it to still fold properly and end up where it was supposed to be, the fact that that one amino acid wasn't there would have a mild effect but probably wouldn't be all that severe.

The problem is it gums up the works. Something about that particular phenylalanine that is supposed to be right there at position 508. If it's not there, this complicated process of getting this whole thing assembled in the origami way gets hung up, and it doesn't finish the job. And parts of it that should have come together in a nice formed, globular way get all strung out over to the side, and it gets stuck in the machinery that is supposed to produce that protein in the right place at the right time, and it doesn't work.

Krulwich: But if you could figure out how to properly re-fold the chemistry, then you would cure someone with cystic fibrosis, or at least make it so mild that they wouldn't die.

Collins: There are several drugs already under development that aim to do exactly this, to sort of cushion that molecule and help it to fold properly, even in circumstances where it normally wouldn't. We know, by the way, if you just lower the temperature, it will fold okay. Something about being at body temperature is just a little too warm for this unstable protein with this mutation in it to find the right shape, and if you could just cool it down a little bit -- not so easy to do -- you could probably convince it to end up properly folded.

Krulwich: Two final questions about this. Why is this such an attractive business? It seems like everybody we talk to who used to be in the gene business seems to want to get into the protein business.

Collins: Well, again, I think the proteins are the players that carry out the action. If you really want to understand a disease, if you want to develop a therapy -- whether it's replacing the protein or developing some small molecule that is going to interact with that protein to nudge it along in a way you want to -- you've got to understand how the protein works.

And, ideally, you want to know all about it. You want to know what is its structure is in three dimensions, what's its active site, what other proteins does it bump into, where is it in the cell, what are the kinetics or the reaction that it's involved in -- all of that stuff. That's proteomics, that's what everybody is all excited about.

We sort of have the genome and at least a pretty darn good working draft. That `s great. That allows you to predict what all the proteins are because of this correspondence. Now it's kind of time to move on and figure out how to use that information therapeutically, and the proteins are key to that.

Krulwich: Is that another way of saying that we have found -- in fact, you have helped find the genetic mistake that leads to cystic fibrosis, but all these years have passed and we can't fix the gene, so let's just go upstairs and see if we can fix the chemistry that the gene creates? Is that basically the deal?

Collins: I think for every disease you're going to see people pursuing therapeutic ideas in two different directions. One will be to understand what is wrong with the gene and try to directly fix it. That is the strategy that we commonly call "gene therapy." The other is to understand how the gene works, what protein it makes, what's wrong with that protein in somebody with the disease, and how to tweak it so that it works after all, even though it wasn't quite designed right.

And for me, I don't care which of those pathways works for a particular disease, whether it's Alzheimer's disease or heart disease or schizophrenia. I just hope one of them does and does so quickly. And I think the betting odds in many situations are that this pathway that works through an understanding of the protein may get us there more quickly than the very difficult problem of how you fix a gene itself when it's not spelled right, and how do you deliver the nice neat correction of that problem, which has clearly been a big challenge for research.

Krulwich: So as hard as it may be to learn the art of folding, you might discover how to be a better folder, you might be able to become a gene fixer. That's maybe a dumb way of saying it. But if it's caught between here and there, if there it's called something that helps the patient -- cystic fibrosis --

Collins: Yeah, that's where we want to get to.

Krulwich: So did you get a little frustrated after a while when you knew the name of the cause but there was no way to fix the cause?

Collins: Well, actually when we discovered the cystic fibrosis gene in 1989, I didn't give much hope that this would have therapeutic consequences for 20 years. It just didn't seem, when we stared at this molecule, that it was obvious even what direction to go in. So actually I think we've made pretty good progress, but not good enough. We still have not seen this disease cured or even particularly benefitted by all of this wonderful molecular biology. You have to still treat it pretty much the way it was 10 years ago.

But that is going to change. There're now about a dozen drugs in the pipeline that are based upon a molecular understanding of this gene and its protein products. And some of the protein therapies are to try to encourage the molecule to fold correctly and some of the others are actually -- now that we understand what it does -- to try to get another protein that might be able to take over the function here, sort of take up the slack, to work a little harder to compensate for its partner, the cystic fibrosis protein, that isn't doing its job. I would call that a protein-based therapy as well. It's using a different strategy, but it still depends on understanding precisely what the protein problem is.

Krulwich: And could you venture a guess as to how long it might be before something interesting happens on the protein side with CF?

Collins: Oh, I'd say a lot of interesting things have happened on the protein side.

Krulwich: Well, let me ask you better, more specifically. Where you could pop something or inhale something and then you would not have as bad a disease?

Collins: Well, it's always risky to predict timetables. I mean, I am encouraged that there are a number of these drugs already in Phase 1 trials to try to see whether they're toxic and whether they might have a hint of benefit. That's a long pathway. The CF Foundation has also put together, I think, a very ambitious program using this thing called "combinatorial chemistry," where you can screen tens of thousands of possible compounds at a rapid pace to try to find others that tweak the cystic fibrosis protein in the right direction in a high throughput kind of analysis.

Krulwich: You mean, I can get tugboats, sort of folding tugboats, that basically shove things together or move them apart, that sort of thing?

Collins: See, the cell has its own system of tugboats to sort of keep things from getting out of line. They're called things like "chaperones," because they are chaperoning along this protein so it doesn't misbehave or get drunk on Saturday night. Sort of make sure that that protein goes to where it's supposed to and doesn't deviate from the straight and narrow.

So there's a system there already, but maybe it needs a little help. And if you had the ability to screen a very large number of possible small molecules you might find one that sort of steps in at just the right moment when the chaperone was looking the other way, or had a particularly difficult client that night and had it all turn out okay after all. Got everybody home safe.

Krulwich: New subject and final subject. The movie. Has there been a movie that has turned your eye a little bit? That raises some of the more interesting ethical questions for the whole genome project?

Collins: Well, I actually think GATTACA for me was the movie that really raised some interesting points. I watched that move three times.

Krulwich: Which movie?

Collins: This movie called GATTACA that was in the theater for about a week and a half. I think I and about four other people went to see it. It was a little over the top, as any movie would be that deals with science fiction. But it was provocative. It portrayed a society where genetic determinism had basically run wild. And society had given up all their civil rights in the belief that genetics was able to predict everything about you in a very precise way.

Krulwich: Let me go to the scene -- in the gripping opening scene a baby is born, and they quickly put some kind of device on his heel, pull out a piece of blood. Then you watch one globe of blood drop. And it looks like that globe is fixed. Apparently everything that that blood can tell you will happen. At least that's the impression the movie gives. And this nurse starts calling off the precise probabilities of one horrible thing after another. Is that conceivably an accurate reading of what might happen in 100 years?

Collins: Well, it's like a good Hollywood depiction ought to be. It has a kernel of truth to it to sort of draw you in, but it's also overstated in order to make a good yarn. So, yes, a drop of blood contains within it enough cells to be able to look at the DNA sequence of you or me or a newborn baby. So it would be realistic, if you decided it was ethical, to actually do this kind of report card analysis of a newborn. In a matter of another 10 or 20 years that will be feasible.

Krulwich: Wait a second. I notice that she is making predictions about near-term diseases and problems, and she even tells the mother and dad when the baby is going to die and of what.

Collins: That's the part that is really off the wall. Because, in fact, these kinds of genetic predictions, while they will be possible for common diseases like heart disease or diabetes, they will be very squishy. They will be able to say this person's risk of getting heart disease at age 70 is two and a half times higher than the average. But there will still be a good chance that they won't get it. So these will be risk factors couched in statistical terms.

And the notion, as the movie portrays, that some of these will be 99 percent -- no way, unless you're talking about something like Huntington's disease, where the misspelling of the gene is almost certainly going to lead to a certain result. But that's quite unusual.

Krulwich: She says -- she gets this sort of sad little look on her face, the nurse does. She says, "Heart disease, 90-something percent." That's the one where you think, "Uh-huh."

Collins: Genetic predictions, except for strongly inherited disorders like sickle cell anemia or Huntington's disease, will not be of that sort. And most of us, when we get this information, will be getting squishy statistical information that says risk is a little higher, risk is a little lower, but you still might not get this or you might. We're going to have to deal with that ambiguity.

And why is it ambiguous? Well, because genetics is not deterministic. There's the environment; it's a big deal. There is free will; that's a big deal, too, depending on what you decide to do with your own body and your own diet and whether you smoke. It would seem immediately obvious from those arguments, which I think most people accept, that it would be ridiculous for somebody to be able to look at your DNA and say, "Oh, you're going to die at age 40."

That's impossible to accept because we understand that it isn't how it works. Look at identical twins, our sort of favorite example of nature and how it interacts with nurture. When you get to know them, they're different people. When you look at their medical records, they're not identical. Yet their DNA is. So clearly that kind of prediction that you see in the movies is going to be couched in all sorts of uncertainties and ambiguities, and I'm just as glad.

Krulwich: But the learning to live with an environment like this, or learning to live with report cards, is something that's going to happen, right?

Collins: Yes.

Krulwich: Because this little boy and this little boy's parents -- because in the next scene the little boy falls and you see the anxious look on the mom's face. Every fall has more meaning to someone who is predicted to be about to have an illness. This "probably" diagnosis is scary. It's very hard to adjust to.

Collins: Yeah. So let's step away from the movie and consider the reality. Ten years from now I think you and I as adults -- I would say probably not as children; I don't think that's such a great idea, most other people agree -- but as adults we could find out what our susceptibilities are for a dozen different conditions: cancer, diabetes, etc.

Would that be useful information? Well, in some instances you bet it would. If you have an intervention -- by lifestyle or diet or medical surveillance -- you could reduce the risk of dying of a terrible disease? Yeah, that sounds like a good thing.

But we have to keep cognizant of the ways in which that might be a bad thing for some people if it does become a self-fulfilling prophecy, if somebody who is bound to have a higher risk begins to think of themselves as already ill, even though they're fine. Or, more diabolically --

Krulwich: Don't you know a lot of people like that; people, who, if there's a cloud in the sky, they know they just live in the shadow?

Collins: People come in all sorts of flavors. And I'm sure genetic information will play into whatever dynamics that individual already possesses about how they view themselves as healthy or not so healthy. And we'll have to deal with that, that we may be, in fact, adding to the burden that people who already feel a little shaky about their condition of health will now have more evidence to feel shaky, and that's something to wrestle with.

And those people, clearly, should never have this information imposed upon them if they don't want it. They should figure out whether this is something they want to carry around or not. I think most people will probably say "No thanks" to the kind of information that doesn't get attached to an intervention that allows you to do something. Doing something gives you a sense of control. Okay, it's this. I can make it lower than this.

Krulwich: But what about Glenn Close? Glenn Close, who is not related to this baby or related to the parent, says, "Sorry, we're not paying your bills because you introduced an expensive human being into the world, a human being who we knew was going to have trouble. I don't know who you are, Mr. and Mrs. X, but you have dome something antisocial."

Collins: This is a serious issue, maybe the most serious one we've talked about. We do have a promising future here in terms of the applications of genetics towards individualized preventive medicine that can keep us all healthy and allow us to live longer, and that's a good thing.

But if that information, as in the movie, gets used to say, "I'm sorry. Your health insurance has been canceled because you're too big a risk anymore," or, "You have to take this test because we, your insurer or your employer, need to know whether we want to support you anymore," then we've done something terrible.

This is a fixable problem. It requires effective federal legislation. I'm sorry, there ought to be a law, and there should already be a law.

Krulwich: So all the villains in this little selection here, Glenn Close is the one you really don't like?

Collins: I think the scenario of predictive genetic information being used to discriminate against people is truly scary and one that we ought to put an end to and we can't.

Krulwich: What about the genetic chemist, that slightly smily guy. He says, "All right. Do you want a boy or do you want a girl?" And then he said, very matter-of-factly, "And we've taken care of all those things that might make your child less than above average." This is a Garrison Keillor world. All the children apparently have to be above average.

That seems to me almost extremely likely to happen, because what parent wouldn't want to introduce a child that wouldn't have -- at least be where all the other kids could be?

Collins: That for me was, in fact, the most compelling theme in the movie, because it's the closest to a real situation that is right in front of us, or almost in front of us. The technology that's being described there is, in fact, quite realistic. It's called "pre-implantation diagnosis," as part of in vitro fertilization. It's being used to try to prevent the birth of a child with a terrible genetic disorder, like, say, Tay Sachs disease. But it allows you to carry out a diagnostic procedure on multiple embryos and then decide which one you want to reimplant.

That has already been somewhat in the news with the case of Adam Nash, the little boy who was the product of such a screening procedure, and who is both free of the disease that his sister has but also a good match for his sister and has provided her with the transplant which she needed. That has certainly caused a lot of ethicists to look at this and go, "Wait a minute. Did we just start down that famous slippery slope?" Because this process has up until now been limited to the use of trying to prevent a terrible disease. But now we've actually created a child partly for another purpose --to provide a donor.

Krulwich: You saw that look in the mom's face. You know, she looks at her husband and says, "Aren't we supposed to say something, because we don't want a perfect baby; we wanted a baby that was without any horrible diseases. We were going to stop there." And this fellow has just rolled right past that, as if, well, any caring parent would feel this way.

Collins: That's why the scenario is chilling. You can sort of see how the dynamic might develop and how parents might in fact feel that in order to demonstrate that they are good parents, that they do care about their kids, that they are going to put them in the best schools and give them music lessons. But it's also not just a possibility but their obligation to make sure that they have the best mix of DNA.

And then there is this wonderful line in there about, "We're not doing anything unnatural here. We're just taking the best of you." There's no new introduction of genetic material in this process; it is simply screening a large number of embryos that that couple might produce and trying to pick the one that will have the most desirable characteristics.

Now, here again there is an element of Hollywood science fiction here. Because the number of embryos that could actually be produced is not huge, and the ability, therefore, to select for 10 different things that would all be optimal in the same one, well, you go through the probabilistic calculations, and it isn't going to work very well. And, furthermore, it's not going to work very well because the whole premise of the movie about determinism is wrong.

So if they're promising this couple that this child is going to be a musical prodigy and is going to do all these other wonderful things, the kid may be a sullen adolescent who takes drugs. You can't actually count upon the DNA being that definitive. So when you begin to go down this pathway of the designer baby, the designer part of it is going to extremely imperfect. And that alone may make this not very appealing to most couples. Just the same it behooves us --

Krulwich: Meaning this is no guarantee here?

Collins: Right, no guarantee. You will not be able to sign a document that says, "I'm going to bring this one back if it doesn't turn out the way I want." And a lot of the things that couples might be most interested in asking about, like intelligence or athletic ability or physical attractiveness, whatever that is, are not going to be very nicely predictable by this kind of technology.

So the good news might be that this attempt to use this technology for enhancement of characteristics is so flawed in terms of its effectiveness that it never really takes off. And then we don't end up with the scene in the movie. But at the same time I don't think you can entirely count on that, at least not in a big way.

And so I do think we are at the point now where we ought to begin to consider whether society has an interest in placing some sort of limits on the use of genetics for enhancement. I think society in general has smiled upon the use of genetics for preventing terrible diseases; that's part of a long tradition, that's what medicine is about, that's what almost every culture believes in. Alleviating suffering is a good thing.

But when you begin to blur that boundary and move into an arena of making your kids genetically different than they would have been by the usual roll of the dice in a way that enhances their performance in some way, that starts to make most of us uneasy.

Krulwich: At the same time if it's a specific question about your baby, it's hard to imagine that there could be a law that says, "You can't have the best baby you can have. We won't allow you" -- this word "best," assuming that such a thing is possible. It's very hard to imagine people not quickly running off to Curacao or the Antilles to get that baby.

Collins: I agree with you. We have a collision course here between the usual principles that couples are the only people who ought to be involved in deciding about their reproductive decisions and their outcomes, and society's interest in not having a particular branch of science run off in a direction that most people in the society are uncomfortable with.

There are consequences if we turn our head and completely look the other way. The technology won't be available to everybody. It will probably be available to those with lots of resources. You can imagine --

Krulwich: Money, you mean?

Collins: Money, yes, money. And education and access to sophisticated medical technology. And this might further distance the haves from the have nots. It might further add to the ways in which people look at disabled individuals as though it was their fault, which very much came across in the movie. There are consequences for just saying, "Oh, well, let's see how this turns out." So I don't think we can afford to say, "It's inevitable. We better just turn our head And walk away." I think we really have to engage in this debate.

Interviews: Collins | Lander | Venter


Photo: WGBH/NOVA.

Watch the Program Here | Our Genetic Future (A Survey)
Manipulating Genes: How Much is Too Much? | Understanding Heredity
Explore a Stretch of Code | Nature vs Nurture Revisited
Sequence for Yourself | Journey into DNA | Meet the Decoders
Resources | Update to Program | Teacher's Guide | Transcript
Site Map | Cracking the Code of Life Home

Editor's Picks | Previous Sites | Join Us/E-mail | TV/Web Schedule | About NOVA
Watch NOVAs online | Teachers | Site Map | Shop | Search | To Print
PBS Online | NOVA Online | WGBH

© | Updated April 2001
/wgbh/nova/genome/textindex.html /wgbh/nova/genome/