 To make a gene that bounds that radioactive molecule more tightly. Yeast is a natural model organism and it worked great. He said, your mother has a fatal disease. It's a genetic disease. By linkage and collecting families, you could in fact find disease genes. Her boy are very quickly recognized commercial possibilities in recombinant DNA. Amid all the turmoil over the technique, he tried to interest several backers in this possibility. At first he failed, but then suddenly success came out of the blue and a new industry was born. The fall of 1976, I get a phone call from a young fellow by the name of Bob Swanson. He says, I've been reading about this recombinant DNA technology. He says, I think it's ready to be commercialized. And I said, yes, I think so. And he says, well, I think I can get some money. And I want to start a company. He says, do you think we could get together and talk about this? And so I said, okay. And so he comes to my lab on a Friday afternoon. And his fellow walks in, not much hair, very young. He's 29 years old, got a suit and a tie. Man, he really looks good. And everybody's, you know, people in my lab are, you know, they're looking at him. You know, who is this guy? So he comes into my office and we sit down and we talked about this. And I just sketched out the idea of putting a company together. Swanson and Boyer's conversation quickly led in April 1976 to the formation of a new company. They called it Genentech for genetic engineering technology. And then in September 1978, at a press conference, they announced that Genentech, using recombinant techniques, had successfully, and for the first time, produced human insulin. The founding of Genentech was key because it brought together the strengths of the academic community with the strengths of the financial community to really move this technology forward, make a difference in human medicine. And it did. I later worked on the set of genes involved in that metabolism or lactose to try to find the control product, the repressor, for those genes. The question is, what is the repressor? What is it chemically? I mean, you could say it's a protein that recognizes DNA. Could it be another piece of DNA or even an RNA? It was considered. We isolated that operator region as a DNA fragment by taking all the DNA to the cell, putting the repressor molecule on it, destroying all the rest of the DNA, letting the repressor molecule protect a small piece of DNA just under it, and then isolating that and looking at it. To look at it, we could do a chemical experiment where we could say, here's a piece of DNA, it's about 20 bases long, a little piece of double-stranded DNA. What does it look like? That's the problem of DNA sequencing immediately. This is back in the early 70s. It's impossible actually at that time to sequence the DNA directly. We tried Alan Maxim and I then spent two years working out the sequence of the 20 bases of DNA. Well, there was a lot of technological development, the development of electrophoretic methods for separating DNA. Of course, out of that came the sequencing, the ability to do sequencing. I, a year or so later, began to do a set of experiments that tried to ask in detail how did the protein touch the operator. Those experiments actually led to a DNA sequencing method because the discovery we made was that we could identify, we could decipher not just which bases the protein touched by these chemical modifications, but we could, in fact, the pattern of breakage to the DNA by these modifications was so sharp and so distinct that we could recognize which bases were which in the sequence. That led to our discovery of a DNA sequencing method. Alan Maxim and I developed a chemical sequencing method which was extremely rapid. At the same time, Fred Sanger and Coulson discovered a enzymatic method that also was extremely rapid. Both methods have, in fact, the same intrinsic speed. They depend ultimately on the same conceptual trick and the conceptual trick is to identify the position of a base in DNA by its distance from some reference point. Now, you can find genes in the fruit fly using breeding experiments, but you can't use breeding experiments with people. But then, at a genetics meeting in Alta, Utah, David Botstein was inspired to propose a method by which you actually might be able to map the location of genes in the human genome. What happened was this fellow Kravitz was talking about linkage of hemochromatosis to HLA. They had a theory and they did the usual human genetics thing. In those days, linkage in humans was almost impossible to do. One of the few markers, maybe a dozen, that was useful was the HLA because it's so polymorphic. They made a theory that hemochromatosis is recessive and linked to HLA. The molecular biologists were having none of this because there was all this mathematics and statistics. I actually understood what they were trying to do and was trying to explain to people that HLA was just a marker. And I remember saying something like, if you had polymorphisms all over the genome, you could map anything. And I suddenly realized that, well, we do have polymorphisms all over the genome with the recombinant DNA. We could see them with southern blots. And Davis was in the room and he heard that and he realized also right away that I had said something that actually was feasible. The polymorphisms that David Botstein had in mind are restriction fragment length polymorphisms called riflips for short. These are just fragments of human DNA that vary in length from person to person. Now, if a disease gene occurs near a riflip in a person's genome, then the riflip can be used as a telltale sign for the presence of that gene. We knew that the entire human genetics field was missing this capacity. That if that capacity were there, we could find all those disease genes, cystic fibrosis and Huntington's disease and all those things. So we had a workshop with David Botstein, Ray White, a large number of other people from the Utah group. And there was a lot of conversation about could you really find these little variations to look for a human gene? And in the hall at MIT, David Hausmann was the one into the hall and David Botstein was the other end of the hall. And everybody was discussing riflips, everybody was discussing restriction fragment length polymorphisms. So they asked us what we thought and we said, well, you know, your odds are really lousy of finding it without the map. With the map we can find anything and that turned out to be true. Without the map, you know, it depends on dumb luck. David Hausmann said, look, you know, why don't we just look for a disease gene and maybe we really will get lucky. Well, dumb luck is what they had. And the eighth marker I think it was that they looked at is in fact linked to Huntington's disease gene. And they said, we got it, we got it, we found it, you know, we found it. We got near the marker, this is it. We got the marker, you know, we found the genes on the top of chromosome 4. You know, I think it really revolutionized the way people thought about what you could do because up to that point, really, even with the paper being published in 1980, by 1983, nobody thought you could do it. So I found myself standing at the bedside of a young nurse in her early 20s who had terrible lung disease and had been sort of a questionable diagnosis for a long time and during that admission we established what she really had with cystic fibrosis. It was clear there was a gene involved. It was also clear that there was no clue what that gene might be or even how you would go about beginning to find it. The really important thing about linkage mapping is that it allowed one to take a disease about which one knew only that it was inherited, that is phenotype. You only knew which of, let's say, five children had the disease and which didn't. And you could use that fact alone to locate a position on a chromosome by its proximity to a landmark and therefore know where to look for the disease gene. And that's how cystic fibrosis was found by Francis Collins and Lapchichoy. By that time, you know, this is now 1982, 1983, the Botstein's idea of how you would use markers to try to map a disease gene was beginning to look like it could actually work. So I approached Lapchichoy. He was a wonderful scientist and very hardworking in Toronto and proposed the idea that we merge our lab's efforts on CF. His approach was complementary to ours. We were doing the jumping thing. He was doing the walking thing. We could jump, then they could walk from where we had landed while we jumped to the next place. We could put together the whole stretch of DNA in an incredibly complicated, laborious way, but it was the best you could do in 1987. And then in the spring of 1989, data began to emerge that one of the genes we had landed on looked pretty interesting. So the defining moment for me was in New Haven at a meeting of the human genome mapping workshops. Lapchichoy and I were both attending the meeting. We thought we had something. He had a fax machine set up in his dorm room. We were all staying in the Yale dorms. Mine was next door. On the other side of him was our chief competitor. The walls were very thin and we repaired back to Lapchichoy's room after one evening session to see whether our labs had generated anything interesting that day. And here it came, this very large tally. And it got to the point where you could see the numbers of unaffected people was getting quite large and the number of affected people were getting quite large. And it became convincing, overwhelmingly so, that evening, to me, this had to be it. This could not be a tease. This was the real thing. It was the first time that a gene for a human disease had been identified by this purely positional approach.