“No one in the mainstream scientific community now takes selection literally.”—Eugene Koonin in conversation with me
“They should have invited Eugene Koonin,” Canadian biochemist Larry Moran told me in the hallway during a break at the November Royal Society “new trends” in evolution conference in London —somewhat exasperated by the proceedings. And I agreed. But Eugene Koonin doesn’t quite see it that way, as he revealed during our recent conversation.
Eugene Koonin is a consummate scientist, Leader of the Evolutionary Genomics Group at the National Center for Biotechnology Information, NIH, who sees science scandals and controversy as unproductive. He prefers the lab to the battlefield and is regarded as one of the most qualitative and prolific thinkers in science today (H-index 196; author of more than 600 scientific papers). [Feb. 2021 Note: Koonin’s H-index now 212.]
In our interview that follows, Koonin sorts out any confusion over his recent statements about the importance of population genetics—part of the Modern Synthesis—and his 2009 statement: “Not to mince words the Modern Synthesis is gone.“ He agrees with Richard Lewontin that the term natural selection is metaphorical and goes further noting, “No one in the mainstream community now takes selection literally.” And he also agrees with the Woese-Goldenfeld perspective about biology as “the new condensed matter physics,” although Koonin thinks biology is still en route there.
Eugene Koonin is the author of several books, among them The Logic of Chance: The Nature and Origin of Biological Evolution; and Sequence—Evolution—Function:Computational Approaches in Comparative Genomics (with Michael Galperin). He is also founder and editor-in-chief of the journal Biology Direct. In 2002, Koonin and NCBI colleague Kira Makarova identified the genetic region now known as CRISPR-Cas.
I spoke by phone with Eugene Koonin at his lab in Bethesda, Maryland.
Suzan Mazur: Do you have any thoughts about the recent Royal Society meeting on new trends in evolutionary biology?
Eugene Koonin: Yes. Perhaps there is a need to make some comment. I was quite unhappy reading at and around this public evolutionary meeting at the Royal Society. Frankly, I think that the less sensationalism, the less controversy brought into these discussions the better. It’s really important that we try at all costs to do normal science rather than some sort of scandalous activity. I was happy about one thing, though—that I was not there and was not directly involved.
Suzan Mazur: Why is it so difficult to pull together the most compelling ideas in evolutionary biology and come up with an approximate understanding of how it all works? Michael Lynch once told me it was because reaching out to other fields is a “daunting task.” But if scientists across the board won’t come together to give us a coherent understanding of how it all works —however approximate—the public will lose confidence in the science establishment’s ability to deliver. This is already beginning to happen. Would you comment?
Eugene Koonin: First of all, I think the public may not have much to lose in terms of confidence in the scientific establishment in this case because the public is already extremely skeptical about the value and the scientific nature of evolutionary biology. It’s not quite that way about science in general but I think largely so when it comes to the study of evolution. Much of the public is poorly informed about it, poorly understands it and is highly skeptical for various reasons. So I would frame the discussion a little differently, in the sense that evolutionary science may not be doing the best possible job to convince the public of the true importance of evolutionary biology. That said, I do believe that a coherent understanding of “how it works” is slowly but steadily emerging in evolutionary biology. However, one has to face the facts: first, it is a slow process, and we are still far from the goal; second, the emerging picture is highly complex and, furthermore, makes little sense without mathematical theory. Thus, communicating modern evolutionary biology (as opposed to deceptively simple antiquated ideas) is indeed a daunting task.
Suzan Mazur: If you were organizing a public evolution summit, what discoveries in biology would you showcase?
Eugene Koonin: I would try to focus on two aspects. One is genomics, and in particular, comparative genomics and metagenomics discoveries—all this comes under the wide umbrella of genomics. That’s one. The other is the existence of solid theory in evolutionary biology. I’ll elaborate on both aspects.
The first aspect, genomics, has in roughly the last 25 years completely transformed the ability to investigate, assess and measure evolutionary processes. All our conclusions on the course of evolution until the advent of genomics had been indirect. It’s remarkable how many of these conclusions and findings remain relevant, but the fact is that all our ways to peer into the evolutionary process and evolutionary past had previously been indirect.
Genomics now provides us windows into the evolutionary past by which we can compare directly the DNA and protein sequences from a rapidly widening range of organisms and thereby make solid conclusions about evolution.
Suzan Mazur: Are you saying this is the top discovery in evolutionary biology in the last 50 years?
Eugene Koonin: The word “discovery” may not apply quite directly here. It’s a transformation of the whole science, which is based on a variety of discoveries. The very approach to evolutionary studies has changed completely. Not only the fact of evolution itself but the existence of deep evolutionary connections between different domains of life—to be concrete—evolutionary connections between, let us say, mammals, such as humans, and prokaryotes, bacteria and archaea, have become indisputable. These findings make questioning not only the reality of evolution but the evolutionary unity of all life on earth completely ridiculous and outside of the field of rational discourse.
Then to be more specific, I would probably showcase the advances of metagenomics—you know, the genomic revolution continues in the sense that now through metagenomics scientists are able to obtain a less and less biased picture of the diversity and evolution of life on Earth. It’s becoming not so unrealistic to think about something approaching a complete picture of the evolutionary history of life.
And then I would showcase something very specific. That is, the latest discovery of the particular group of archaea that was the direct ancestor of eukaryotes. And in this case, “discovery” is the right word.
There is a necessity to bring to the broader audience of biologists and lay public Mike Lynch’s reformulation of the principles of genomics in terms of population genetics. Paraphrasing the famous pronouncement of Theodosius Dobzhansky, one of the founding fathers of the Modern Synthesis [“Nothing in biology makes sense except in the light of evolution.”], Lynch wrote in one of his papers: “Nothing in evolution makes sense except in the sense of population genetics.” That is absolutely true. The details of population genetics theory are difficult to explain even to biologists who are not specially trained, yet we have to communicate these ideas to a broader audience, including the lay public, and in qualitative terms.
Suzan Mazur: How much of the research in your lab is bench experiments and how much is computer modeling?
Eugene Koonin: That’s easy, 100% of the research in my lab is computational, not necessarily modeling, but 100% is done by computer and 0% is done experimentally. Of course, we constantly collaborate with experimental laboratories.
Suzan Mazur: In a 2009 paper of yours commenting on the 150th anniversary of the Origin of Species you make the following statement: “So, not to mince words, the Modern Synthesis is gone.”
Yet in your recent BioMed Central paper you write that it’s time for biologists to start paying attention to population genetics because of advances in functional genomes. But population biology IS part of the Modern Synthesis. So your current position has some in the science community confused. Would you talk about the evolution of your thinking about evolution and begin by how you define “gene” and “genome”?
Eugene Koonin: Such confusion makes one wish, at least for a moment, they never made such general statements aimed at a mass audience, yet I think such generalizations are necessary. There isn’t really much change in my thinking. There isn’t any dramatic change let alone a turn-around in my thinking. Population genetics is a mathematical framework that is essential for building evolutionary theory but it is not the theory itself. The Modern Synthesis does employ that framework and is a correct theory but only for a narrow range of evolutionary processes in certain groups of organisms. It is quite a typical situation in science, actually.
Suzan Mazur: There’s also a lot of confusion in the evolutionary biology community about what a gene is. For instance, Jim Shapiro says he doesn’t think in terms of genes as entities. He thinks in terms of systems all the way down.
Eugene Koonin: This is a completely different level of discussion. Let’s try to separate it, whatever. I know exactly what Jim Shapiro said and a lot of people say. It’s just a translation into a different language, from a somewhat different viewpoint. I do not disagree, genomes are dynamic systems evolving in space and time not static collections of genes. But it is also OK to view them as entities, information storage devices. These viewpoints are complementary.
Coming back to the evolution of my thinking from 2009 to 2016, which really hasn’t been much. Quite frankly, if I were writing what I wrote back in 2009, I would have been even more cautious and non-combative than I was then. I don’t think I was ever really bombastic. But I would have been even less demonstrative and maybe I would not have written that the Modern Synthesis is gone.
Suzan Mazur: I think your paper in 2009 does sort of leave the door open for the paper that you just published in BioMed Central.
Eugene Koonin: Absolutely, all the doors were open. I would not say that it’s [Modern Synthesis] gone just like that. It has to be understood in context. I think now any actively working scientist in evolutionary biology probably realizes that the Modern Synthesis or neo-Darwinism, or whatever the name is, is insufficient in the post-genomic era. This is a set of concepts that is insufficient for understanding the entirety of evolution. It doesn’t mean it’s wrong. It’s only becoming wrong if someone claims that they need nothing past the concepts in the Modern Synthesis.
Suzan Mazur: Again, there are complaints in the evolution science community that nothing ever seems to get solved. You’ve commented in the BMC paper that there continues to be a parade of just-so narratives and that “if biology is to evolve into a “hard” science with a solid theoretical core, it must be based on null models, no other path is known.” You note further that null models are standard in physics but not in biology. Would you say more, beginning with your definition of “null model.”
Eugene Koonin: Sure. In any field, null model is the simplest explanation of the available data that does not violate physical laws. Good and sensible scientific practice in physics but also in other sciences. Scientists first come up with the simplest rational explanation of the available data and then see if anything in the data refutes that explanation and requires a more complex model. And so on and so forth.
Suzan Mazur: Nigel Goldenfeld in recent years referred to biology as the “new condensed matter physics.”
Eugene Koonin: Yes. He wrote a paper with the late Carl Woese where they expressed this, and I agree. Maybe with a caveat. I would rather say biology has to become the new condensed matter physics.
Suzan Mazur: At the November Royal Society public evolution meeting mentioned above, Sir Pat Bateson cautioned about the overuse of the metaphor of natural selection. And Richard Lewontin has famously said in the New York Review of Books that Darwin never meant the metaphorical term to be taken literally by generations of scientists. You keep natural selection in your most recent BMC paper and identify a family of selection terms: “weak selection,” “purifying selection,” “positive selection,” “local selection,” and “global selection.” Aren’t these all metaphorical as well and contrary to your interest in seeing biology “evolve into a ‘hard’ science”?
Eugene Koonin: Well. Yes, these are metaphorical. From Darwin to this day. I also agree with Lewontin, Darwin did not mean natural selection to be taken literally. But we have to be, I guess, a little more specific about what it means to take natural selection or any kind of selection literally. It means, one would assume, the existence of a selecting agent. Perhaps making all these parallels between natural selection and artificial selection, the way Darwin does in his book, could be somewhat dangerous because in artificial selection there is someone who is selecting, even if unconsciously. In that respect, the evolutionary process is very different in nature where nothing is there to actually select. Darwin certainly realized this and wrote more precisely of “survival of the fittest.” In modern evolutionary biology, it is sometimes “random survival” but the key point remains the same: organisms survive and leave progeny differentially. I think it is quite alright to denote some forms of differential survival selection, metaphorically. And there is no confusion here, within mainstream thinking. No one in the mainstream scientific community now takes selection literally.
Suzan Mazur: You also say in the BMC paper: “Counterintuitive as this might seem, evolutionary reconstruction in my laboratory clearly indicates that the ancestral state in most major groups of eukaryotes and apparently the last common eukaryotic ancestor had an intron density close to that in extant animals.” You note that introns persist in eukaryotes because introns invaded their genomes as mobile elements early on and that selection was too weak to get rid of them. You also say “the substantial majority” of introns harbor no detectable gene.
What is the significance of this observation? And would you define, in this case, what you mean by intron because you cite two groups of introns in your October table of defined virus terms. Thank you for that paper, by the way—it’s very useful—the paper in Studies in History and Philosophy of Science Part C.
Eugene Koonin: Thank you. I appreciate that. What was said there in the virus paper?
Suzan Mazur: You provide an extensive table of defined virus terms, and you identify two groups of introns: Group I and Group II.
Eugene Koonin: We are going into technicalities here, so just very briefly. Prokaryotes also have genetic elements that are called introns but they’re very, very different from eukaryotic introns. Prokaryotic introns are more like mobile elements, self-splicing introns, unlike the eukaryotic introns that just sit there and wait to be excised by the spliceosome.
Prokaryotic introns are active. They have the machinery to excise themselves and even to move to a different location. There are two classes of such self-splicing introns, Group I and Group II, but the distinctions between these groups are important only for those who study these things.
The point for the general reader is that the eukaryotic introns evolved from the Group II self-splicing introns, which invaded the genomes of early eukaryotes and then lost their mobility.
Suzan Mazur: Would you touch on the possible importance of stem-loop RNA in origin and evolution of life?
Eugene Koonin: It’s a bit of an unexpected turn. All RNAs contain stems and loops, all RNAs that exist in life forms, in organisms are stem and loop structures.
Eugene Koonin: I cannot right now comment on the specific statements of Luis. I simply don’t remember them. Sounds very generic as long as one believes in the primordial RNA world, in some form. Yes, within the RNA world model, stem and loop structures are essential. But random stems and loops do not form the right structure, they cannot have ribozyme activity let alone complex ribozyme activity. So they are only starting material for pre-biological evolution, they do not solve any problems by themselves
Suzan Mazur: In a presentation last year in Tokyo at ELSI (Earth-Life Science Institute) on the emergence of the biosphere, Uppsala University scientist Ajith Harish pointed out that “advances in our understanding of protein evolution indicate that tertiary structures of proteins are the molecular fossils of evolution while coding sequences are transients.”
Harish also says the Universal Common Ancestor of the contemporary Tree of Life (ToL) “is distinct from any specific modern descendant,” that the Universal Common Ancestor was not the first cell lineage and that the modern ToL is the crown of a “recently” rerooted tree, that “bottlenecked survivors of an environmental collapse, which preceded the flourishing of the modern crown, seeded the current phylogenetic tree.”
Harish concludes that the “new data raises questions about traditional hypotheses based on sequence-based gene trees as well as divergence time estimates based on limited information in gene sequences,” noting further that “there are so far no identifiable ‘universal’ viral genes that are common to viruses such as the ubiquitous cellular genes.”
Would you like to comment on this?
Eugene Koonin: The short answer is no, I do not want to comment on that, because it’s impossible to make any responsible comment on a long and complex quote like this unless I’ve heard the lecture (or much better yet, read the paper). There are a variety of things on which I would agree (for example, that protein structures are more conserved than sequences which is common knowledge) and a variety of things on which I cannot immediately agree. But the bottom line is I did not hear the lecture.
Suzan Mazur: At the same ELSI meeting, Hiromi Saito from Osaka University questioned whether the common ancestor of bacteria had a cell wall noting, “many bacteria can transform themselves to a cell-wall-deficient state” called an L-form. Do you have any thoughts about that?
Eugene Koonin: I know very well what L-forms are, in particular with respect to their simple cell division mechanism. And this is an interesting possibility when we think about early evolution of cells. The modern L-forms obviously are derived, and comparative genomics tells us that the last common ancestor of bacteria probably did have a cell wall. Wall-less forms might have been important in evolution of cells but at an even earlier stage.
Suzan Mazur: Would you like to make a final point?
Eugene Koonin: Yes. I would like to come back to this issue of the Modern Synthesis, population genetics theory and the like because it is true that population genetics theory is part of the Modern Synthesis. And that is great. That is part of the power of the concept and why it remains quite relevant in explanations of microevolution but also an important part of the new evolutionary biology. That’s what I wanted to convey in the BMC paper, that population genetics theory (in its modernized form because it too has not remained static over 50 years) has to be systematically applied in evolutionary genomics, which is the new mainstream of evolutionary biology. Indeed, it’s changed dramatically over the last 25 years, and as previously mentioned, has completely transformed the ability to investigate, assess and measure evolutionary processes. The modern version of population genetics theory (it too has not remained static over 50 years) has to be actively, constantly and systematically applied to our understanding of genome evolution. That is too often not the case.
The whole of Mike Lynch’s work on this, his talks, papers and books are of paramount importance, even if I sometimes disagree with Mike on specific issues. The foundation Mike Lynch laid for modern evolutionary genomics cannot be reasonably disputed and is of huge importance.