Darwin and the web of life
January 30th 2009 10:23
This article in New Scientist by Graham Lawton, titled "Why Darwin was wrong about the tree of life", has sparked a minor uproar - or rather, this was achieved by the front cover of the issue in which it appeared. Readers have felt that the cover title is misleading and may provide grist for the creationist mill. In large letters, it reads "Darwin was wrong", though with a subtitle "Cutting down the tree of life". The article itself, which should give no comfort to creationists, is fascinating and informative, though I personally think that it may itself engender some confusion. I don't think it quite measures up to the message conveyed in either title. Its focus is upon whether most of the life on Earth has evolved in a branching tree fashion or whether it's more like a web, with horizontal gene transfer (as opposed to vertical gene transfer to progeny) taking place on a large scale. This is an interesting and worthwhile question, and it is the centre of much current research and controversy. But back to why I think the article may be misleading (some have cited it as an example of journalistic hype, and personally I do feel it was somewhat irresponsible. But I won't push the matter further here):
(I have basically paraphrased a message I posted on richarddawkins.net)
A few years ago, National Geographic asked the question "Was Darwin Wrong" on the front cover of an issue of theirs. They gave the answer "No". Who's right? The National Geographic article beautifully illustrated a number of Darwin's ideas that have been vindicated by a wealth of evidence. They took the question to mean whether evolution occurs and whether we can discern relatedness between species based upon shared morphological and embryological features as well as the evidence in the fossil record showing transitional forms. Surely, that they could answer their question in the negative with clear cases of evidence and thereby affirm some basic premises of Darwinian evolution (or rather, evolution per se, because the evidence mainly tells us that evolution has occurred, not so much the actual mechanisms involved. However, this too is changing and now geneticists can detect the occurrence of natural selection and other processes by looking at tell-tale patterns in genomes) suggests that New Scientist may, if anything, be focusing too tightly on a particular aspect of Darwin's theory and taking that as its "essence".
There are five senses I would like to briefly touch upon to convey the idea that this may be apt to give a misleading impression to those not versed in evolutionary theory. The first is that Darwin, like any other great scientist, was bound to be corrected or updated sooner or later about virtually anything he said. He did, after all, write The Origin 150 years ago. As may have mentioned, he didn't know about genes, much less could he have known about horizontal gene transfer and hence of evolutionary alternatives to the tree of life. So in this first sense in which the heading is misleading, one could reply "So what?" As Olivia Judson has said, calling modern day evolutionary biology "Darwinism" is a little like calling modern day aeronautics Wrightianism. One can still talk sensibly about Darwinism because modern evolutionary biology still shares fundamental points of contact with Darwin's original insight, but clearly the game has moved since his time, and there are many things that enrich and at the same time complicate those insights.
The second sense is related to the first. Clearly, species are know to be generated from ancestral stock, and a process like cladogenesis can account for this. So to a first approximation, Darwin is still right. He worked with what he knew, and his hypothesis was entirely sensible. We can think of the tree as being a universal potentiality that is - in many cases - actually complicated and even overridden. It happens to be an empirical fact that many lineages appear to be more web-like than tree-like, but the tree concept is still a useful first-principles concept that is still seen in many cases anyway.
Thirdly, we know of a whole bunch of molecular sequences (like redundant gene and protein sequences but also pseudogenes and transposons) that give convergent phylogenies, with probabilities against this being due to hierarchically nested genealogical affinity infinitesimally small. This at least is clear in many eukaryotes, though of course the picture will be greatly complicated by HGT in bacteria and archaea, and even some eukaryotes. In these cases, it will be difficult to find many mutually convergent phylogenies because the lineages in question will have been constantly irrigated by genetic material from all over the place, blurring any tree, perhaps making it useless. I think the article may have given the impression that, due to HGT, molecular phylogenetics is dead because some species are so amalgamated with genes from different lineages that these neat convergences cannot be realised. I think they should have qualified this a bit more, lest it give anyone the impression that cladograms are outdated in their entirety. But it's still more subtle than this. Convergent phylogenies tell us about vertical gene transfer from one lineage to another. It's still conceivable that, for example, two species that are shown to be related by virtue of convergent redundancies in separate genes happen to differ in a great many of their other genes because they were each being irrigated "side-ways on" (through HGT), acquiring genes and such from other species. In that case we would say that the species are each more "closely related" to other species than they are to each other when it comes to those genes.
Fourthly, and related to point three, when we're talking about genes, there is still undoubtedly a tree of life. One can imagine a phylogeny of a gene, starting from a common ancestor and onto its divergent forms, and whether they've branched off via HGT plus mutation or vertical transmission plus isolation of populations is quite beside the point.
Fifth, none of this has much bearing on Darwin's most important and ground-shaking insight: cumulative natural selection. HGT is a way of presenting a lineage with genetic variability, but this variability will be acted upon by selection as though it had appeared through any other route. This part of the Darwinian edifice stands outside the debate about a tree of life or a web of life. So in this sense, Darwin has not be shown to have been wrong, and this is worth labouring since it is all to easy to confuse one aspect of a theory with the theory as a whole. Even if the tree of life were to be utterly and completely swept aside, we would still have a plausible, coherent mechanism to explain how complexity and adaptive features are generated.
Be that as it may, I really recommend that everyone read the article. It presents another case of how our understanding of life's history is continuously changing, and how much more complex and strange our world is than we could ever have imagined even a few decades ago.
(I have basically paraphrased a message I posted on richarddawkins.net)
A few years ago, National Geographic asked the question "Was Darwin Wrong" on the front cover of an issue of theirs. They gave the answer "No". Who's right? The National Geographic article beautifully illustrated a number of Darwin's ideas that have been vindicated by a wealth of evidence. They took the question to mean whether evolution occurs and whether we can discern relatedness between species based upon shared morphological and embryological features as well as the evidence in the fossil record showing transitional forms. Surely, that they could answer their question in the negative with clear cases of evidence and thereby affirm some basic premises of Darwinian evolution (or rather, evolution per se, because the evidence mainly tells us that evolution has occurred, not so much the actual mechanisms involved. However, this too is changing and now geneticists can detect the occurrence of natural selection and other processes by looking at tell-tale patterns in genomes) suggests that New Scientist may, if anything, be focusing too tightly on a particular aspect of Darwin's theory and taking that as its "essence".
There are five senses I would like to briefly touch upon to convey the idea that this may be apt to give a misleading impression to those not versed in evolutionary theory. The first is that Darwin, like any other great scientist, was bound to be corrected or updated sooner or later about virtually anything he said. He did, after all, write The Origin 150 years ago. As may have mentioned, he didn't know about genes, much less could he have known about horizontal gene transfer and hence of evolutionary alternatives to the tree of life. So in this first sense in which the heading is misleading, one could reply "So what?" As Olivia Judson has said, calling modern day evolutionary biology "Darwinism" is a little like calling modern day aeronautics Wrightianism. One can still talk sensibly about Darwinism because modern evolutionary biology still shares fundamental points of contact with Darwin's original insight, but clearly the game has moved since his time, and there are many things that enrich and at the same time complicate those insights.
The second sense is related to the first. Clearly, species are know to be generated from ancestral stock, and a process like cladogenesis can account for this. So to a first approximation, Darwin is still right. He worked with what he knew, and his hypothesis was entirely sensible. We can think of the tree as being a universal potentiality that is - in many cases - actually complicated and even overridden. It happens to be an empirical fact that many lineages appear to be more web-like than tree-like, but the tree concept is still a useful first-principles concept that is still seen in many cases anyway.
Thirdly, we know of a whole bunch of molecular sequences (like redundant gene and protein sequences but also pseudogenes and transposons) that give convergent phylogenies, with probabilities against this being due to hierarchically nested genealogical affinity infinitesimally small. This at least is clear in many eukaryotes, though of course the picture will be greatly complicated by HGT in bacteria and archaea, and even some eukaryotes. In these cases, it will be difficult to find many mutually convergent phylogenies because the lineages in question will have been constantly irrigated by genetic material from all over the place, blurring any tree, perhaps making it useless. I think the article may have given the impression that, due to HGT, molecular phylogenetics is dead because some species are so amalgamated with genes from different lineages that these neat convergences cannot be realised. I think they should have qualified this a bit more, lest it give anyone the impression that cladograms are outdated in their entirety. But it's still more subtle than this. Convergent phylogenies tell us about vertical gene transfer from one lineage to another. It's still conceivable that, for example, two species that are shown to be related by virtue of convergent redundancies in separate genes happen to differ in a great many of their other genes because they were each being irrigated "side-ways on" (through HGT), acquiring genes and such from other species. In that case we would say that the species are each more "closely related" to other species than they are to each other when it comes to those genes.
Fourthly, and related to point three, when we're talking about genes, there is still undoubtedly a tree of life. One can imagine a phylogeny of a gene, starting from a common ancestor and onto its divergent forms, and whether they've branched off via HGT plus mutation or vertical transmission plus isolation of populations is quite beside the point.
Fifth, none of this has much bearing on Darwin's most important and ground-shaking insight: cumulative natural selection. HGT is a way of presenting a lineage with genetic variability, but this variability will be acted upon by selection as though it had appeared through any other route. This part of the Darwinian edifice stands outside the debate about a tree of life or a web of life. So in this sense, Darwin has not be shown to have been wrong, and this is worth labouring since it is all to easy to confuse one aspect of a theory with the theory as a whole. Even if the tree of life were to be utterly and completely swept aside, we would still have a plausible, coherent mechanism to explain how complexity and adaptive features are generated.
Be that as it may, I really recommend that everyone read the article. It presents another case of how our understanding of life's history is continuously changing, and how much more complex and strange our world is than we could ever have imagined even a few decades ago.
| 39 |
| Vote |
