Evolutionary biology
Sep. 17th, 2004 04:25 pm![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
tarigwaemirAd Mundo Exteriore,
Yesterday, instead of finishing off data analysis, I had a very interesting conversation with the postdoc at my lab on evolutionary theory. He'd found this fascinating paper in Nature (citation: Lenski, et al., The evolution of complex features, Nature 423, 139-144) that studied selection in an artificial computer environment. They looked at "digital organisms" or pieces of code that had the ability to replicate, like viruses only with the added possibility of mutations (that is, less than perfect copying to allow for variations) and let them loose in a closed system. The computer environment was set up to have limited energy resources, and the ability to perform complex functions would be rewarded by more energy being allotted to that particular "organism". Thus, a digital equivalent of selection for fitness.
Now all of the above is old hat by now, but they were trying to model how organisms would be able to evolve very complex structures. As Stephen Jay Gould mentioned numerous times in his various Natural History columns and books, the rather radical concept in Darwin's Origin of Species was not that life evolved but rather that a slow and negative process like natural selection could act as a creative force. Darwin himself raised the famous example of the eye, and the question of how such a perfect and intricate structure could evolve by small variations, in Origin, and he answered with the concept of gradualism: the accumulation of small changes over wide stretches of geological time.
Lenski et al. looked to see how long it would take their initial "organism" to evolve a particularly complex function, which required logic functions that the original code did not have. (In order to drive selection in this direction, the ability to perform this function, EQU, would be rewarded with the most energy units. Ability to perform simpler functions would also be rewarded, based on their level of complexity.) The code in fact was given only one logic function, "NAND", which could potentially mutate and recombine to provide all of the operators requisite to performing EQU. The simplest and most efficient way to evolve to EQU required sixteen steps. Since selection depends on random variation, none of the final organisms that managed to evolve EQU followed such a simple path. On the other hand, the paths they did follow proved to be quire informative.
The researchers found that the evolution of EQU depends on the evolution of simpler functions, but there was no unique "intermediate stage" that all organisms passed through in order to achieve EQU. I quote: "Following any particular path is extremely unlikely, but the complex function evolved with a high probability, implying a very large number of potential paths." In other words, evolution really is convergent--there are many paths to the same destination. Furthermore, the difference between the organisms who could perform EQU and their parents who could not only consisted of a few mutations--however the ability to perform EQU depended on multiple other mutations that had already accumulated, often for quite different purposes. In other words, epistatic interactions are the norm. The transition from genotype to phenotype is a highly networked process. Most interesting of all, in some lineages, a deleterious mutation was a necessary precursor to the final mutation that resulted in ability to perform EQU. Also, they noticed that sometimes simpler functions (that increased fitness) were lost in order to evolve more complex functions. Thus the path towards maximum fitness via increased complexity is not linear--it has setbacks and tangents like any form of history.
In some ways, this experiment supports the idea of punctuated equilibrium--radical changes in phenotype occurs quickly, taking advantage of the previously accumulated genetic variation (which may have been responsible for much less drastic phenotypic variation). Of course, you can easily retort it's just gradualism from a different perspective. In any case, it's a really beautiful picture of how natural selection works and reflects our current knowledge of how interconnected and complex our genome is. A single mutation can be neutral, or it can have quite drastic effects on phenotype, due to its nonadditive interactions with other mutations or genetic variation already present. Note the word nonadditive. The whole world of biology functions on the concept of emergent properties. The whole is always more the sum of its parts; complexity evolves from simple components. (I'm waxing Hofstadter, aren't I?)
Anyway, towards the end of the paper, I was particularly struck by this line: "As a consequence, selection acts on organisms rather than directly on their genes." The sentence refers, I think, to the whole "selfish gene" debate. Since neither the postdoc or I knew much about "selfish gene" theory (other than that Richard Dawkins was its loudest proponent), we turned to the papers and to Stephen Jay Gould for a better understanding. The "selfish gene" theory seems, at first, mostly to be a matter of semantics and perhaps conceptual aestheticism. It claims that the genes are exclusively the units of selection, and that organisms are merely the vehicles for which they efficiently reproduce themselves. We read some quotes from Dawkins that wax rather poetic on the theme--rather offputting to a scientific perspective actually--but actually, Gould did a good job of outlining the motivations behind such a perspective. He's an excellent historian in that respect. I'm going to outline what he writes below:
Firstly, the "selfish gene" had its heyday when reductionism was the fad in biology (quite possibly the era when all the physicists and chemists turned into molecular biologists). The era isn't quite over yet, and even now, there's a pretty heavy emphasis on a reductionist approach to biology. We try to explain everything in terms of molecules these days. Of course, now we have a better respect for the whole concept of emergent properties and try to study how the different levels of genome, cell, organism and population all interact, but the tendency is there. Thus, to reduce evolution to the genetic level, to explain selection in terms of genes rather than organisms, made sense from a philosophical point of view.
Secondly, there's also the issue of primacy. I think most people agree that the genetic molecule preceded the cell, and in a population of self-replicating molecules (ah, "RNA world" theory, my one true love), selection does occur. (It's really elegant how logical selection is. All you need is replication, source of variation and differential fitness.) This has all been tested in vitro, by the way. Gould points out that taking this model into account, one could easily consider cells as an "adaptation" of certain genetic molecules (after all, genetic molecules surrounded by a membrane is surely going to out-compete nucleic acid molecules that are vulnerable to decomposition), and then expostulate the rest of life as further elaborate ways for genomes to replicate themselves. He points out the fallacy with this conceptual framework is again that emergent properties are ignored. If organisms really were no more than their genetic and protein constituents, then the "selfish gene" theory would be legitimate. But organization means that once more the whole is not the sum of its parts. (Gosh, I sound like a broken record.)
Thirdly, and perhaps the least "philosophical" and the most "scientific" argument is that selection works on the principle that beneficial variations are amplified in the population because the individuals with those variations are able to pass them down to their offspring. However, when we consider sexual reproduction, only half the genes are inherited and randomly too--the fitness of the parents do not completely assure the fitness of the offspring. Thus, selection should be considered to occur at the genetic level, not the organismal level. One might be tempted to conclude that it's still a matter of semantics, and in many ways it actually is.
So why should we care whether people agree or disagree with "selfish gene" theory or not? In the end, it doesn't really affect our ability to explain evolutionary history and biological processes, does it? From a historical or philosophical standpoint, it is of course very important, but to the average experimental scientist (who often consider the theorists to be wild and crazy kooks), it seems to be useless nitpicking. So I thought about why "selfish gene" or even refuting "selfish gene"--personally I think it sound rather extreme and don't agree with it much at all, although since I basically learned evolutionary theory through Stephen Jay Gould, I'm more than a little biased--is important. And I remembered that on the first day of Bio. Sci. 50 last year, Professor William Fixsen began his introductory lecture with what was basically a paraphrase of selfish gene theory. He said that we are simply vehicles for our DNA to reproduce itself efficiently. A humbling thought, you have to admit, and in retrospect, I have to disagree with the sentiment. But the very presence of "selfish gene" affects the way professors present and articulate ideas. The truth is, semantics do matter. Language affects the way we conceptualize, and experimental science requires us to construct meaning from our data. (A point which is often forgotten. I think one of the things I adore about my lab is that no matter how much we get mired in data analysis, someone's always around to go on high flights of theoretical fancy, to keep us thinking about how it all fits in.) The narrow reductionist viewpoint of "selfish gene" theory may seem a matter of wordplay, but it can seriously hamper our ability to see the larger picture.
Gould by the way doesn't dismiss the gene as a unit of selection--he just considers it one of the many hierarchical units of selection. In fact, one of Gould's major theoretical contributions to evolutionary theory is articulating this idea that selection works at many different levels in the biological hierarchy: gene, organism, population, species, even phyla. He considers the organism to be perhaps the major unit, but by no means the only one, and argues that we must look at how selection at different levels interact in order to explain many evolutionary phenomena. There were quite a few really interesting examples of "suborganismic selection"--it's a hard concept to wrap your mind around, but there were some really interesting cases--and I meant to go through my favorite ones, but my fingers hurt. I think I'll save that summary for tomorrow.
One last comment: Lenski, the primary author of the paper I mentioned at the beginning, seems to have published several other interesting papers on experiments with selection. He's mostly worked with E. coli, which reproduces quickly enough for microevolutionary change to be observed. There was one we saw where he used gene arrays to see how levels of expression were changed--that is so awesome--and another where he studied selection and epistasis. If you're like me and get a thrill of near-physical pleasure at seeing clever experiments in evolutionary biology like these, you should look his name up on PubMed. ^_^
I know, I'm such a biology geek. Our lab by the way is applying for a grant for a joint project on hybrids and how two separate transcription networks are integrated in a viable hybrid organism, which makes me happy. (Hybrids are quite possibly a major source of speciation, particularly in plants, which can handle polyploidy much better than we can.)
I bet several of you are bored to tears by now, although personally I can't see why because this is really exciting stuff if you sit and read through it.
...Tari
Yesterday, instead of finishing off data analysis, I had a very interesting conversation with the postdoc at my lab on evolutionary theory. He'd found this fascinating paper in Nature (citation: Lenski, et al., The evolution of complex features, Nature 423, 139-144) that studied selection in an artificial computer environment. They looked at "digital organisms" or pieces of code that had the ability to replicate, like viruses only with the added possibility of mutations (that is, less than perfect copying to allow for variations) and let them loose in a closed system. The computer environment was set up to have limited energy resources, and the ability to perform complex functions would be rewarded by more energy being allotted to that particular "organism". Thus, a digital equivalent of selection for fitness.
Now all of the above is old hat by now, but they were trying to model how organisms would be able to evolve very complex structures. As Stephen Jay Gould mentioned numerous times in his various Natural History columns and books, the rather radical concept in Darwin's Origin of Species was not that life evolved but rather that a slow and negative process like natural selection could act as a creative force. Darwin himself raised the famous example of the eye, and the question of how such a perfect and intricate structure could evolve by small variations, in Origin, and he answered with the concept of gradualism: the accumulation of small changes over wide stretches of geological time.
Lenski et al. looked to see how long it would take their initial "organism" to evolve a particularly complex function, which required logic functions that the original code did not have. (In order to drive selection in this direction, the ability to perform this function, EQU, would be rewarded with the most energy units. Ability to perform simpler functions would also be rewarded, based on their level of complexity.) The code in fact was given only one logic function, "NAND", which could potentially mutate and recombine to provide all of the operators requisite to performing EQU. The simplest and most efficient way to evolve to EQU required sixteen steps. Since selection depends on random variation, none of the final organisms that managed to evolve EQU followed such a simple path. On the other hand, the paths they did follow proved to be quire informative.
The researchers found that the evolution of EQU depends on the evolution of simpler functions, but there was no unique "intermediate stage" that all organisms passed through in order to achieve EQU. I quote: "Following any particular path is extremely unlikely, but the complex function evolved with a high probability, implying a very large number of potential paths." In other words, evolution really is convergent--there are many paths to the same destination. Furthermore, the difference between the organisms who could perform EQU and their parents who could not only consisted of a few mutations--however the ability to perform EQU depended on multiple other mutations that had already accumulated, often for quite different purposes. In other words, epistatic interactions are the norm. The transition from genotype to phenotype is a highly networked process. Most interesting of all, in some lineages, a deleterious mutation was a necessary precursor to the final mutation that resulted in ability to perform EQU. Also, they noticed that sometimes simpler functions (that increased fitness) were lost in order to evolve more complex functions. Thus the path towards maximum fitness via increased complexity is not linear--it has setbacks and tangents like any form of history.
In some ways, this experiment supports the idea of punctuated equilibrium--radical changes in phenotype occurs quickly, taking advantage of the previously accumulated genetic variation (which may have been responsible for much less drastic phenotypic variation). Of course, you can easily retort it's just gradualism from a different perspective. In any case, it's a really beautiful picture of how natural selection works and reflects our current knowledge of how interconnected and complex our genome is. A single mutation can be neutral, or it can have quite drastic effects on phenotype, due to its nonadditive interactions with other mutations or genetic variation already present. Note the word nonadditive. The whole world of biology functions on the concept of emergent properties. The whole is always more the sum of its parts; complexity evolves from simple components. (I'm waxing Hofstadter, aren't I?)
Anyway, towards the end of the paper, I was particularly struck by this line: "As a consequence, selection acts on organisms rather than directly on their genes." The sentence refers, I think, to the whole "selfish gene" debate. Since neither the postdoc or I knew much about "selfish gene" theory (other than that Richard Dawkins was its loudest proponent), we turned to the papers and to Stephen Jay Gould for a better understanding. The "selfish gene" theory seems, at first, mostly to be a matter of semantics and perhaps conceptual aestheticism. It claims that the genes are exclusively the units of selection, and that organisms are merely the vehicles for which they efficiently reproduce themselves. We read some quotes from Dawkins that wax rather poetic on the theme--rather offputting to a scientific perspective actually--but actually, Gould did a good job of outlining the motivations behind such a perspective. He's an excellent historian in that respect. I'm going to outline what he writes below:
Firstly, the "selfish gene" had its heyday when reductionism was the fad in biology (quite possibly the era when all the physicists and chemists turned into molecular biologists). The era isn't quite over yet, and even now, there's a pretty heavy emphasis on a reductionist approach to biology. We try to explain everything in terms of molecules these days. Of course, now we have a better respect for the whole concept of emergent properties and try to study how the different levels of genome, cell, organism and population all interact, but the tendency is there. Thus, to reduce evolution to the genetic level, to explain selection in terms of genes rather than organisms, made sense from a philosophical point of view.
Secondly, there's also the issue of primacy. I think most people agree that the genetic molecule preceded the cell, and in a population of self-replicating molecules (ah, "RNA world" theory, my one true love), selection does occur. (It's really elegant how logical selection is. All you need is replication, source of variation and differential fitness.) This has all been tested in vitro, by the way. Gould points out that taking this model into account, one could easily consider cells as an "adaptation" of certain genetic molecules (after all, genetic molecules surrounded by a membrane is surely going to out-compete nucleic acid molecules that are vulnerable to decomposition), and then expostulate the rest of life as further elaborate ways for genomes to replicate themselves. He points out the fallacy with this conceptual framework is again that emergent properties are ignored. If organisms really were no more than their genetic and protein constituents, then the "selfish gene" theory would be legitimate. But organization means that once more the whole is not the sum of its parts. (Gosh, I sound like a broken record.)
Thirdly, and perhaps the least "philosophical" and the most "scientific" argument is that selection works on the principle that beneficial variations are amplified in the population because the individuals with those variations are able to pass them down to their offspring. However, when we consider sexual reproduction, only half the genes are inherited and randomly too--the fitness of the parents do not completely assure the fitness of the offspring. Thus, selection should be considered to occur at the genetic level, not the organismal level. One might be tempted to conclude that it's still a matter of semantics, and in many ways it actually is.
So why should we care whether people agree or disagree with "selfish gene" theory or not? In the end, it doesn't really affect our ability to explain evolutionary history and biological processes, does it? From a historical or philosophical standpoint, it is of course very important, but to the average experimental scientist (who often consider the theorists to be wild and crazy kooks), it seems to be useless nitpicking. So I thought about why "selfish gene" or even refuting "selfish gene"--personally I think it sound rather extreme and don't agree with it much at all, although since I basically learned evolutionary theory through Stephen Jay Gould, I'm more than a little biased--is important. And I remembered that on the first day of Bio. Sci. 50 last year, Professor William Fixsen began his introductory lecture with what was basically a paraphrase of selfish gene theory. He said that we are simply vehicles for our DNA to reproduce itself efficiently. A humbling thought, you have to admit, and in retrospect, I have to disagree with the sentiment. But the very presence of "selfish gene" affects the way professors present and articulate ideas. The truth is, semantics do matter. Language affects the way we conceptualize, and experimental science requires us to construct meaning from our data. (A point which is often forgotten. I think one of the things I adore about my lab is that no matter how much we get mired in data analysis, someone's always around to go on high flights of theoretical fancy, to keep us thinking about how it all fits in.) The narrow reductionist viewpoint of "selfish gene" theory may seem a matter of wordplay, but it can seriously hamper our ability to see the larger picture.
Gould by the way doesn't dismiss the gene as a unit of selection--he just considers it one of the many hierarchical units of selection. In fact, one of Gould's major theoretical contributions to evolutionary theory is articulating this idea that selection works at many different levels in the biological hierarchy: gene, organism, population, species, even phyla. He considers the organism to be perhaps the major unit, but by no means the only one, and argues that we must look at how selection at different levels interact in order to explain many evolutionary phenomena. There were quite a few really interesting examples of "suborganismic selection"--it's a hard concept to wrap your mind around, but there were some really interesting cases--and I meant to go through my favorite ones, but my fingers hurt. I think I'll save that summary for tomorrow.
One last comment: Lenski, the primary author of the paper I mentioned at the beginning, seems to have published several other interesting papers on experiments with selection. He's mostly worked with E. coli, which reproduces quickly enough for microevolutionary change to be observed. There was one we saw where he used gene arrays to see how levels of expression were changed--that is so awesome--and another where he studied selection and epistasis. If you're like me and get a thrill of near-physical pleasure at seeing clever experiments in evolutionary biology like these, you should look his name up on PubMed. ^_^
I know, I'm such a biology geek. Our lab by the way is applying for a grant for a joint project on hybrids and how two separate transcription networks are integrated in a viable hybrid organism, which makes me happy. (Hybrids are quite possibly a major source of speciation, particularly in plants, which can handle polyploidy much better than we can.)
I bet several of you are bored to tears by now, although personally I can't see why because this is really exciting stuff if you sit and read through it.
...Tari
![[identity profile]](https://www.dreamwidth.org/img/silk/identity/openid.png){kind=small}
(no subject)
Date: 2004-09-18 09:41 pm (UTC)the bulk of it kind of hit my brain and died though. this means i'm going in for a second helping tomorrow because... you're right. it IS, well, interesting stuff [considering how sleepy i am right now, it'd take a LOT to get me excited]. =)
p.s. i've been meaning to catch up on some bio reading as of late... and you seem the type to point me in the right direction. i managed to find a copy of gould's "the panda's thumb" and... for some reason i'm blanking out. granted, reading time is precious little before sat iis and college apps... but i think i'll try to spoil myself nonetheless. =)
(no subject)
Date: 2004-09-18 09:59 pm (UTC)I have to admit that I haven't read a Gould book cover to cover yet. I usually pick up his essay anthologies and read the ones I find interesting. I say, don't feel bad if you blank out because later you find yourself remembering a lot more than you think you absorbed. ^_^ That always happens to me.
I've pretty much learned my evolutionary biology exclusively from Gould (you know, our AP Bio textbook followed the Gould school of thought pretty closely, which isn't surprising since one of the contributors, Vrba, wrote lots of papers with Gould), but if you want to be a better scholar than me and read the other side, Richard Dawkins' The Selfish Gene and The Blind Watchmaker is the way to go. From what quotes I've seen, he has beautiful, extravagant prose--a little too extravagant for science, I think--and may be fun to read. I've been meaning to get around to reading them for ages, but as you said, reading time is so precious. Heh. ^_^;;
Good luck with your apps and SAT IIs! Fighting!
...Tari