03 December 2015

Mikhail Gelfand: "Science benefits by its very existence"


The famous Russian bioinformatician Mikhail Sergeevich Gelfand told how to understand biology using a computer, answered the question whether it is possible to clone a mammoth, and explained how the interaction of science and society is similar to the relationship of aphids and ants. 


Mikhail Sergeevich Gelfand graduated from the Moscow State University Mehmat. Doctor of Biological Sciences, Professor of the Faculty of Bioengineering and Bioinformatics of Moscow State University, Head of the Laboratory and Deputy Director for Science at the A.A. Harkevich Institute of Information Transmission Problems of the Russian Academy of Sciences, member of the European Academy, laureate of the A.A. Baev Prize. Research interests: comparative genomics, metagenomics, metabolic reconstruction and functional annotation of genes and proteins, search for regulatory signals, evolution of metabolic pathways and regulatory systems, gene expression and alternative splicing, spatial structure of chromatin. Active public figure, member of the Public Council of the Ministry of Education and Science, one of the founders of the Dissernet, co-founder and member of the Board of the Educational Fund "Evolution". Deputy. editor–in-chief of the newspaper "Troitsky variant - science".

– They said about you that you do not develop methods, do not write programs, but try to understand biology with the help of a computer. How's that?

– This statement has three aspects. First, surprisingly, it turns out that just looking at the letters of DNA decoding, comparing their sequences in different organisms, it is possible to give answers to traditional biological questions. For example: what does this protein do; in which cases the gene turns on and works, and in which cases it does not; we know someone in the cell performs a specific function – who is it? The second aspect is newer, it has only existed for the last ten years. With the help of bioinformatics, you can see how the cell as a whole is arranged. Answer questions: how are all those genes that turn on and off under the same conditions related to each other? Who regulates them? They are already less traditional, because molecular biology was a reductionist science most of the time – the whole was disassembled into bricks and they were studied – and now we want to look at the whole "house". And here bioinformatics is no longer an additional tool, but a completely integral part of the whole process. When people come up with an experiment, they figure out from the very beginning how to process these results. If this is not done, then real tragedies happen. I have seen several such situations where researchers from the very beginning did not think a little, spent a lot of effort, time and money and received data with which nothing can be done. 

But this bioinformatics is useful, and molecular evolution is interesting. There is a famous saying of Theodosius Dobrzhansky: "Nothing in biology makes sense except in the light of evolution." Like many banal truths, it is very deep. There is no supreme genetic engineer, there is no plan according to which everything was done and which we need to understand. It turns out exactly the opposite: evolution works with what it has. 

– What is currently at the forefront of modern bioinformatics?

– Firstly, today there is just a revolution with data - there are a lot of them. The number of decoded DNA, information about the work of genes, about their interaction, about the structure of the genome is growing rapidly. This allows you to look at some things that were hard to even think about before, because they were just speculative fantasies. Take, for example, the packaging of chromatin in a cell: DNA three meters long is folded into a nucleus only a few microns in size. It turns out that this is due to how genes function: where the packaging is denser, they are silent, where they are slightly looser, they work. If earlier there was practically no experimental data about this, now everything can be simply calculated. 

Secondly, absolutely amazing things happen from the fact that there are many genomes – 20 years ago we had 2-3 of them, and now there are hundreds. On the one hand, you can't work with them using the old methods, you can't touch 100 DNA with your hands, and on the other hand, just by comparing them, you can observe very subtle evolutionary changes. Imagine that you had two lists of the chronicle, differing in large chunks, and then it became 100, and you can trace what each scribe contributed. That's the sort of thing we can do with the genome now. 

Thirdly, it is now possible to study evolution and developmental biology at the same time. We are not like a mouse, although the genes are almost the same. The difference is determined by how they turn on and off at the earliest stages of development. Now, thanks to the analysis of their work in one individual cell, we can look into embryogenesis and find out when and where we "split up" with the mouse. Apparently, there will be a breakthrough in understanding development from the point of view of comparative biology and evolution, precisely at the molecular level. 

Another important area of modern bioinformatics is the study of diseases, primarily cancer. In fact, the cells of our body are different. A classic example: one of them, as a result of a change in the genome, will "go crazy" and become cancerous, a tumor will grow out of it, also consisting of different cells: some will be drug–resistant, the second will have the potential to form metastases, and so on. In order to understand the development of the disease and develop treatment through it, it is necessary to monitor not the fate of the tumor as a whole, but the fate of its individual cells, more precisely, the work of the genes in them. 

Another important area: ancient DNA. Our ideas about how a person happened are changing completely radically. 40,000 years ago, three different lines of humanity roamed Eurasia: Neanderthals, Cro-Magnons and Denisovans. They entered into different relationships with each other, traces of which can be physically seen in the genome. In my opinion, this is terribly cool. No less interesting is the history of the evolution of horses, plants, now there is a wonderful project "1000 genomes of rice".

– You mentioned that the mammoth's DNA was preserved very poorly to clone it. But can we in the future, using the preserved genetic information, "revive" later extinct species (or those that are dying out now)?

Firstly, we do not know how to synthesize DNA of the right length. Secondly, even if it succeeded, it needs to be properly packaged, which modern science does not know how to do and is unlikely to be able to do in our lifetime. There are examples for simpler organisms: you can transplant DNA and turn a bacterium of one species into a bacterium of another, albeit closely related. However, the gap between them and multicellular is enormous, so you can definitely forget about cloning a mammoth. The next problem: in order for the experiment to be successful, the "mother" – the animal from which the clone will be made – must be genetically similar to the sample, otherwise there will be a conflict between her and the embryo. What you can do is do reverse engineering. The mammoth and the Asian elephant had a common ancestor. First, it is necessary to restore it from a modern animal, and from it already – a mammoth. It is important not only what changes have occurred, but also in what order.

– How does bioinformatics and experiment relate? Can it happen that chemistry and biology will almost completely "go into the computer"?

– Of course, the whole experiment will not go into the computer, although it happens in different ways here. On the one hand, when you say that this protein does this, it's important that someone checks. One of the strongest experiences in my life happened when we published a purely theoretical paper where we predicted how certain genes regulate the work of others. The article was more about evolution, but on the way we made a lot of specific predictions collected in a separate table. And then I got to a university in Germany, where, when I came to another laboratory, I talked to a female professor, who looked very strict and severe. She had our work on her desk, open on the table, and there were ticks in some of the cells. I shuddered. The woman noticed my look and reassured me that everything was fine so far. That is, they followed this article and checked everything on the dotted lines. For a theorist, this is quite a strong experience. And in this sense, bioinformatics is a very responsible science. On the other hand, there are predictions in which I am so sure that if an experimenter comes to me and says, "Nothing happened," then I will answer: "The test tubes should have been washed better." Biology will never go into pure theory – after all, it is based on facts. In order to do it in a computer, you must first put the data obtained experimentally there.

– According to you, there is so much genetic information today that it does not have time to fit into the system. As a result, there are some colossal amounts of data, from which only the cream is removed. What do we lose because of this?

– I really have a feeling that in some sense we are not thinking it through, and the available data would allow us to do something much more interesting. Now, if there is new data, a new method, it is necessary to process everything very quickly, because the same thing is happening in the neighboring laboratory, and the question is who will be the first to be published in Nature, and who will be in the bulletin of the Dark–Balkan Culinary College. Science has become very competitive, and it's not God knows how good. The cockroach racing element is very tedious. It seems to me that it would be very interesting to organize such a project: to gather a crowd of bioinformatics and evolutionists and send them to a desert island without the Internet for six months – so that they could calmly think and talk to each other. 

– In your opinion, is the voiced problem a systemic problem of bioinformatics?

– This is a systemic problem of the whole science, which, on the one hand, has become very competitive, on the other hand, there are absolutely delusional ideas that it should benefit. Science benefits by its very existence. The benefit for her is a secondary product that occurs like honeydew in aphids – she poops with sugar syrup, which ants then feed on. Science and society are arranged in a similar way: scientists are engaged in what they are interested in, and society, if it is properly organized, is able to turn the products of this activity into technologies, into medicines, into all sorts of useful things.

– It turns out that Russian society does not know how to do this?

– The problems of Russian science are not problems of science, but of Russia – they are still much more systemic. Although some aspects of this are everywhere, they manifest themselves in different ways. In a democratic society, science competes with healthcare, education, and sports. All these areas should explain to taxpayers why they are needed, since they contain them. As a result, there is a certain inflation of promises – in order to be given money, you have to offer something useful right away. 

It is very difficult to convey the idea to the taxpayer: "I'm going to study number theory, but I don't know if something useful will come out of it." This is not a random example. In the 20s-30s of the last century, there was a wonderful mathematician Godfrey Harold Hardy, who advised his students to study number theory, since it is "the most useless field of mathematics." He was a real English gentleman and believed that science should not benefit. And then all secure transactions on the Internet came out of this area of knowledge – when you put a card into an ATM, it is the theory of numbers that "sits" inside it. 

Science as a kind of human activity has completely proved its usefulness for civilization, but it is very rare for scientists to know which area will "shoot". It is very difficult to convey this idea to the taxpayer, we have to promise some specific things, more and more, and then inflation occurs, and this is bad. Another thing is that it has always been so. Kepler earned money with horoscopes, knowing everything about them perfectly, and for the soul he was engaged in his elliptical orbits; alchemists did interesting things, and the elector was promised a philosopher's stone and gold made of lead. It's just that electors used to be swindled, and now they have to swindle several tens of millions of taxpayers.

In Russia, another problem is that it is not the citizens who need to be liked here, but the authorities, which is much more disgusting. In the first case, there is at least an element of feedback. In the same Europe, in the USA, scientists read popular science lectures, appear on television, in newspapers, because they understand that their budget for the next five years depends on how they explain to society what they are doing. And we need to explain to the secretary of the district committee or some official.

– If little depends on the taxpayer, then why develop scientific popularization in Russia?

– The public demand for people to be told about science, apparently, is still there, and quite strong. I have traveled a lot this year with public lectures, a lot of people come there, and they ask very reasonable questions. How much money should I use to do this? There was a "Dynasty", it was dispersed, but the people who were doing it did not go away, as well as the public inquiry. The Evolution Foundation appeared in the dry balance. We don't have one big sponsor, but it turns out that there are a significant number of medium–level donors – people who are quite successful in business and ready to invest in the popularization of science. And on the other hand, we did crowdfunding on Planeta.ru and in 10 days we collected 800 thousand rubles. This can be considered in some sense a sociological experiment. The answer to the question: is scientific popularization needed in Russia – yes, because there are many people who are ready to give money to make it happen. The situation with the Evolution Foundation makes me optimistic: we have come up with interesting things that can be done, a lot of people come who want to interact and communicate with us, and obviously there is a request for all this to happen. And again – there is no need to explain anything to the authorities.

Diana Khomyakova talked
Photo by Yulia Pozdnyakova

Portal "Eternal youth" http://vechnayamolodost.ru
03.12.2015
Found a typo? Select it and press ctrl + enter Print version