TARTU, Estonia (GenomeWeb) ― Andres Metspalu likes to tell people that life for him began at 40. A 63-year-old professor of biotechnology at the University of Tartu in this northern European country and director of the university's Estonian Genome Center, he entered the scientific field during the waning decades of Soviet rule, when the only journal in the laboratory library was the European Journal of Biochemistry, and he and fellow scientists had to build their own instruments.
"We were completely isolated," Metspalu said during a recent interview with GenomeWeb. "We couldn't buy things from the West," he said. "I designed my own electrophoresis apparatus which was built in the lab workshop."
Since Estonia regained its independence 23 years ago, its trajectory – and Metspalu's – have been quite different. The EU country commenced the Estonian Genome Project in 2001, collecting 52,000 samples by the end of the decade. In 2007, the Estonian Genome Center, which is overseen by the university, was established to conduct active research based on the biobank.
Now, the center is preparing to move toward introducing genomic medicine via the country's online medical data infrastructure. To do that, the center intends to genotype the repository using a customized high-density SNP array. The Estonian Genome Center then hopes to feed medically relevant genetic information into the online medical records system, aiding in national preventative health efforts.
Others have taken note. Earlier this month, Eric Lander, president of The Broad Institute, gave a talk at the center, calling the country a model for implementing genomic medicine. GenomeWeb spoke with Metspalu at the time of Lander's visit. Below is an edited transcript of that interview.
The Estonian Genome Project kicked off a few years after Decode Genetics in Iceland. EGP was a public project, Decode was completely private. When it comes to biobanks, do you think the public model or the private model is a better route?
It depends on what the goal is. For me, yes, I like the public model because I would like to see that genomic medicine will become a part of normal medicine. We would like to provide an additional instrument, genetics, to physicians. Now they have X-ray machines, MRI, ultrasound, whatever. But not genetics for common diseases. Okay, for rare Mendelian diseases, they run tests, but that's a very tiny part of it. But we have common diseases, [such as] cardiovascular disease, type 2 diabetes. It's very much genetics. And we can now stratify the patients according to their genetic risk. We can see the disease risks much earlier and use preventive methods or treat them differently. If you want to make money at the end, make a private biobank, try to develop new drugs or diagnostics, and finally you can make an exit and sell it. But here we have a different attitude. It was our position from day one that we will never sell the biobank. And this is what the Estonian Human Gene Research Act stipulates. We can sell access, we can sell services, we can sell knowledge. If partners want to sequence all the samples, we are happy to work with them. But it cannot happen in Riga or Helsinki or anywhere. It has to happen here.
Have the samples been made available to such partners?
Absolutely. And we have lots of data. Most of the collaborations are based on data. About 25,000 samples have been genotyped. But arrays have been coming and going, and that's 25,000 samples with about six different arrays. Some of these arrays are not very compatible or specific. So now we are in a situation to use this information in medical disease risk and drug response prediction, and for that purpose we are developing a new array and all 52,000 samples will be run on the new array. We have done some good science. We have published in collaboration with other genome centres several hundred papers in the past few years. Now we believe we know enough that we can start personalized medicine here. For those purposes, we will sequence a few thousand people to get the rare variants that are Estonian-specific. We will use one of the commercial arrays that has been developed by the two major companies in the field, but we have to go through a tender process. We would like to supplement this array to put a few hundred thousand more variants on top of it.
When would you commence that project?
This will be a pilot project. There is a road map and plan for the next four years. By September next year we have to have this genotyping ongoing. But it can take eight months, meaning that by spring 2016 it has to be done. Then we will have to use decision support software that will analyze all 52,000 samples and produce reports covering just the major things, a few lines of text at the beginning. It will go to the e-health online system, and when the physician together with the patients open the computer it will say, 'You have an increased risk for glaucoma. You have to go see your eye doctor.' And if the eye doctor says that the risk is high, they will measure your ocular pressure. If it is normal, they might ask you to come back next year, if you are older. Or if you figure out that the pressure is higher, they might prescribe eye drops. We have 28,000 people in Estonia who are blind because of glaucoma. Of course, you can't prevent it for everyone, but if you can prevent 10 percent of them, let's say 3,000 people, then they can work and they can contribute to society, and not get help from society. Or consider coronary heart disease, myocardial infarction. There is strong evidence of a set of SNPs that can predict higher coronary heart disease risk. These people should be taking statins. If we transfer the Finnish data to Estonia, then about 115 people in Estonia are dying each year for the simple reason that they do not get statins at the right time. But it's still the early days of genomic medicine. It will get better as we go along.
Are there certain diseases that are particularly common in the local population?
We are focused on two diseases that are very common — one is cardiovascular disease the other is type 2 diabetes. With cardiovascular diseases, we are number one in Europe. Much of this is related to behavior, but genetics is also important. Lots of drinking, lots of smoking, lots of watching TV and drinking beer. And too little preventative medicine. We have not decided yet about doing cancer, although it is the number two killer. Fifty percent of people are dying from cardiovascular disease, 30 percent are dying from cancer. But cancer is too difficult. You know, I am an optimistic guy, but somatic cancer is rather complicated because we just don't know enough. And now what we see is if we sequence tumors early, we can stratify them. For example, with gastric cancer, we can stratify them into four different diseases, basically. And each of these four diseases can have a different drug. But you have to sequence them and now it is not happening yet at a large scale, this is just research. So it is complicated.
There seems to be an increased interest in serving biobanks as a customer segment. Fluidigm has a SNP panel for DNA quality control, for instance.
Unfortunately, we had to solve most of these problems already 10 years ago. Everything is already in place and we can't change anything for this biobank. We have our quality control systems, storage systems, a biobanking robot for cherry picking samples. I see what is happening, but at this stage everything is working, and we are not going to change anything, but I am keeping my eyes open. When we do the next stage and sample, say, a hundred thousand people, then we need new solutions — new DNA extracting robots, and also the genotyping must be more automatic. We are interested in using the new Affymetrix Axiom arrays. The assay is based on ligation, it's plate based. This is really good. But their oligos are still synthesized the same way. The assay is also better because it is enzyme based, so the signal is better than just hybridization only.
Are you interested in taking part in more population studies through the Estonian Genome Center?
To some extent. For example, we signed a memorandum of understanding with Hanoi Medical University in Vietnam. They would like to establish a biobank, they would like to do population-based analysis and produce a genetic map of Vietnam after sampling 1,000 people. We are ready to go though and provide data as soon as the samples are ready.
What have you learned about local population genetics given your resources?
We are much closer to the other Baltic countries than the Finns. The Gulf of Finland is a barrier between us. And the Alps are also a barrier between Central Europeans and the Italians. But if you take Europe, as a whole, the most important markers which contribute to these types of PCA maps have to do with lactose intolerance, immune system genes, geographic distances, et cetera. It's the more environmentally influenced genes that determine the European genetic map, not so much whether you speak German or French.
How do you see the center in terms of competitiveness?
In certain areas, we are very competitive. Of course, we have only 16 PhDs. But what we are doing in population genomics, in epigenetics, in genomic variation, is close to most of the top guys, but we cover only a small area. I think we are among the first that can implement genomic medicine in the whole country because of e-health system that we have here, which covers everybody. It is a state-organized thing. In very few places can you see the infrastructure to implement personalized medicine at the country-wide level. And putting together e-health, genetics, IT support, and public support for IT solutions, we are at the forefront. The UK, Finland, the Netherlands and several others are also advacing rapidly. But this kind of common database doesn't exist anywhere else. As a genome center, we are quite competitive, I think. Of course our volumes are smaller, but what we do, we do very well.
You were also instrumental in developing arrayed primer extension, or APEX, a microarray-based SNP genotyping approach, in the 1990s. Is APEX still being used?
Some people tell me that I was one of the first who introduced the concept of genotyping arrays to the genetics community for mutation analysis about 20 years ago. APEX is used by [Tartu, Estonia-based] Asper Biotech for certain applications. For some diseases, you have to analyze certain mutations and this method is still best. Asper also has customers all over the globe that are doing interesting things using APEX.