Tell us about how the genetic linkage map got off the ground? What prompted it?

One of the things I like to do is make tools that are of general interest, tools that people can use to try to solve different types of problems. Back in the late 1980s, I had a project to construct a new kind of genetic map using new types of markers that had been identified some time earlier in so-called microsatellites. And so in fact I was very lucky. We submitted this project to the people here at the French Myopathy Association. They had raised a lot of money with a telethon approach —around 50 million dollars, which was a considerable sum in the late 1980s. And they wanted to use that money to start some studies related to the human genome project. So they started several projects, one of which was the physical map led by Daniel Cohen, and another was the genetic-map project that I had in mind. This was in 1990.

  Did you have competition on the genetic map?

Yes, of course. This coincided with the beginning of a U.S. project. We had one advantage, which was that we could operate in a very centralized way. At Généthon, all the money and people and resources could be concentrated in a single facility, whereas in the States there were some fights about who should do the map and so forth. So we had a slight advantage and, of course, we used it. We came out with the first map at the end of 1992.

  How did it influence the human genome project?

It really changed things in the area of mapping disease genes in humans. We saw a spectacular acceleration in the rate of mapping genetic diseases in the early 1990s because of this map. Then we produced two additional versions—one in ’94, and the last in the beginning of ‘96.

  What changed with each new version?

Each version has a higher resolution, which means more markers. The first map had 800 markers. The last one had 5,000.

  What happened to the U.S. competition?

They constructed their map. Their map and ours were quite complementary. They were also using microsatellite markers, but a different kind. So the markers on the two maps were distinct. Their map was half the resolution of ours, so they didn’t go as far as we went, but they had a good map. Their first version came out around 1994.

  You went to work for Genoscope after you finished the third version of the map. What prompted the move, and what were you hoping to accomplish?

This was the time when the various sequencing centers were being established around the States, and at the Sanger Centre in the U.K. There were some centers in Japan and Germany, and we were a little bit late, whereas we were a little bit early for the mapping project. The decision wasn’t made until 1997 to have a sequencing center in France—Genoscope—and I was asked to direct this center because of my contribution to the mapping project.

  So what was Genoscope’s contribution to the genome project?

Since we were, let’s say, a medium-sized center, much smaller than the larger ones, we focused on a single chromosome, which was also a medium-sized chromosome. This was chromosome 14. And we published the sequence of this chromosome two or three years ago.

  What was the single most difficult part of the genome sequencing?

Finishing it. Filling all the gaps and having good annotation, which means having a good interpretation of the sequence. That has to be done using additional data and various informatic tools—different types of programs that make predictions or comparisons, and then you have to decide if those predictions make sense or not. This has been improved lately, but just a few years ago it was still quite a challenge to have an appropriately annotated chromosome.

  What else has Genoscope been doing?

Our other big project was to sequence a fish genome, which was a compact genome, of the puffer fish Tetraodon nigroviridis. [See O. Jaillon, et al., Nature, 431(7011): 946-57, 2004.]

  Why a puffer fish?

Because those fish have very compact genomes. For roughly the same number of genes as in mammals, you just have to sequence about eight times less DNA. The puffer fish genome is around 400 million base pairs, compared to 3 billion in mammals. It’s quite small. When we started this, as I said, we had big difficulties interpreting the human genome. Now we could make comparisons between these puffer fish sequences and human sequences, and that enabled us to identify genes. The first thing we did, in fact—this was at the end of 1999—was to analyze the fraction of the human genome that was available at that time, around 40% of it, and compare it to our puffer fish sequence, which was also incomplete. Just by matching the two sequences we could extrapolate that there would be something like 30,000 genes in the entire human genome. This was much fewer than what was claimed and predicted in those days. This number of 30,000 was confirmed when we had the draft sequence in 2001. In fact, it’s probably a little less; it’s probably around 25,000. But we were much closer to the reality than those people who were claiming that there would be 100,000 genes.

  How did people take your prediction?

They didn’t believe it. We were able to convince the paper’s reviewers, more or less, but they also accepted another paper which claimed something like 80,000 genes. So even the reviewers didn’t want to believe we were right. After that there were still more papers published claiming our estimate was probably wrong and that there would be many more genes in the human genome. Finally, there’s now a general agreement on this rather small number.

  Do you think the genome project, since its completion, has significantly changed the field?

Well, consider, for the moment, that we had to have genome sequences because we had to have inventories of the genes. This was the first goal. Once we had those inventories, we could also identify the genes that are responsible for genetic diseases. For other diseases, the ones we call multifactorial diseases, it’s much more complicated. These are ones in which there are genetic components but also environmental causes. So in those cases, it’s much more difficult to identify predisposing genes or susceptibility genes. For Mendelian diseases, it’s quite easy. Using the sequences has really helped a lot to find those genes. Of course, now that we have this catalog of the human genes, people are asking different questions. They’re asking what are all the genes expressed in different types of cells or tissues. People are now trying to undertake these global approaches, to have a global picture of what’s going on in a cell. I’m still not convinced that this is most urgent thing to do, but everyone is doing it.

  What would you do instead?

Well, these global approaches are useful in a sense, of course. But I don’t think they will really provide a lot of clear pictures of how a cell or an organism is working. We’re still so ignorant about so many of the individual genes and proteins. We don’t know their functions. We probably know the function, or part of the function, of a third of the genes—maybe a little bit more. So I don’t think we can describe in molecular terms what’s going on in a cell, and we don’t completely understand what’s going on in a cell. Even if we have all the different elements, the different parts, we have to know what all those parts are doing. We have to make a lot of progress in cell biology, because there are many types of molecular machines, of small devices, that are made of assemblies of different proteins, and we have to understand this logic, and we can’t do that yet. To me, as long as we don’t understand this logic, I don’t think we can predict anything.

  So when you make decisions on research, do you go with the fashion or with your intuition?

It’s exactly that: fashion or intuition, yes. Our center here, in fact, has two types of activities ongoing. One is to serve the community, and at present most of our sequencing activity is devoted to projects submitted by external investigators. So we do the sequencing for them, and we help them interpret the sequence using bioinformatics. For our own purpose, we’re now concentrating on going back to very old work, to very basic biochemistry. In particular, we’re trying to find new enzyme activities, especially those that are encoded by bacterial genes. So this is one of our goals at the moment. Even in bacteria, we still don’t know the exact function of between 20% and 40% of the genes. Many of those encode just very basic enzymes having to do with the basic metabolism in bacteria. We’re studying that; we’re revisiting bacteria metabolism now that we have the entire genomes sequenced. I think this has to be done, so we’re really trying to complete the missing parts of central metabolism in bacteria.

  How do you see the field of genomics developing over the next five years?

The first thing I have to say is that when you make such predictions, you’re always wrong. And, as I said, personally I’m not involved that much in genomics anymore. Rather, I’m going back to these old questions—although, of course, using genome data. One thing to note, however: there are still plenty of areas in a huge number of species for which we have no known genome sequence yet. We need those sequences, because comparative genomics is very important, and not just in higher invertebrates, for instance, or in mammals. We have to extend this to all the different kingdoms, and there is a lot to do there, a lot to learn. This is all very important. Proteomics and transcriptomics, all those things, will generate a lot of data, but we won’t be able to interpret those data without these other data, as well.

  So is it fair to say that you’re skeptical about the mainstream approach in genetics at the moment?

I am rather skeptical. I’m becoming something of a genome skeptic, or you could say an "omics" skeptic. The genome sequence is great because that information is made of stone, whereas those other data are very soft, with plenty of artifacts. And I don’t think we’re pursuing this in the right way

The human genome sequence is quite reliable and we can do a lot with it. Of course, I think the goal that we had initially when we set out to sequence the human genome was to have the gene catalog in order to find diseases. This is something that is still going on, and it will take a lot of time. We will eventually be able to understand much more about physiology and disease. This is the old approach, though, based on genetics. You have a system and you manipulate one element in the system and see how it reacts. This is what genetics provides, and there’s still a lot to do. Systems biology—the "omics" approach—is great, but premature, because we don’t know enough about the different elements.

	of Jean Weissenbach.