How often do genes move between distantly related species?

April 10, 2015 • 8:45 am

Did you read Matthew’s post on the okapi yesterday? I hope so, because I’m worried, in view of the paucity of comments on science posts, that people are skipping them. Perhaps that just reflects the dearth of things that non-scientist readers have to say. I hope that’s the answer, for it takes about four or five times more work to do a science post than, say, an anti-theist post.

So I’ll try once more today. This is a genetics paper that came out a few weeks ago, but I haven’t had time until now to read it and summarize it. It’s on the phenomenon of horizontal gene transfer (HGT), whereby genes move between the genomes of very distantly related species. Examples include the include the absorption of bacteria by other species (that’s how the mitochondria originated, a theory suggested by Lynn Margulis; and these mitochondria, once free-living bacteria, contribute their genes to the absorbing organism); the transfer of pigment genes from fungi to aphids, which give the aphids a red color that may help hide them; and the transfer of several enzymes from bacteria to insects, which help the insects use new plants as food.

The phenomenon of HGT has been called “non-Darwinian,” since it simply wasn’t envisioned by Darwin or, indeed, even in the early days (ca. 1930-1940) of the “modern synthetic theory” of evolution.  But, contra some critics of evolution, HGT does not invalidate the modern theory of evolution. For the transfer of genes between distantly related species is simply a new source of genetic variation—like “normal” mutations—whose fate is still subject to whether the transferred genes are good or bad for the host. In the case of aphids, for instance, the acquired pigment genes rose in frequency in the pea aphid species by natural selection, but had they been deleterious they would have been eliminated.

This kind of gene transfer can occur in several ways: by eating of one organism by another, and then incorporation of that organism’s DNA into the genome; by infectious transfer of microorganisms followed by the same kind of incorporation; or simply by absorption of microorganisms into the body, as when rotifers rehydrate after they’ve become desiccated.

Infrequent HGT, then, doesn’t kill the theory of evolution, but expands it by showing that “mutations”—the raw material for evolution—can be acquired in a way we didn’t previously suspect.

Now if HGT was very, very common, then it would completely efface the evolutionary relatedness of organisms as seen from their DNA. If Drosophila species, for example, got genes repeatedly from microorganisms, and different fly species got different genes, then one might be completely thrown off by using DNA sequences to determine how related they are. You’d mess up your phylogeny by including horizontally acquired DNA from unrelated species in your tree-making algorithm. This was the basis of the infamous New Scientist cover that read “Where Darwin Went Wrong.” The “wrong” bit was supposedly that Darwin envisioned a branching bush of life, but HGT might mean that the branches were effaced by transfer of DNA from distantly related species, and we wouldn’t have easily defined bushes at all.

Fortunately for evolutionists, HGT isn’t that common—certainly not common enough to prevent us from reconstructing evolutionary relationships, as scientists recently did for families of birds (see here). Cries that “Darwin was wrong!” or “Evolutionary theory is disproven!” are simply wrong. The branching bush of life is still secure, though there are bits of bark that move between the branches.

A new paper in Genome Biology by Alastair Crisp et al. (reference and download below) is the first attempt to systematically find out how much HGT there is between three groups of metazoan organisms (nematodes, flies, and primates) and simpler ones (fungi, microbes, algae). What they did was perform genome scans of DNA sequences of species in all these groups, looking for those gene sequences in nematodes, flies, and humans that were far more similar to sequences in the other species than to more-closely related species of metazoans. For example, they could find a gene in one or several species of fruit flies that was far more similar in sequence to a gene in a bacterium than to any genes in other metazoans. That would imply that that gene had been transferred from bacteria to flies. (Another possibility is that the gene was not the result of HGT, but was present in the common ancestor of all of these species and was simply lost in all the non-fly metazoan species. But the authors used controls to rule out that possibility.)

The authors also divided up the genes that presumably moved by HGT into three classes, A, B, and C, differing in the assurance with which HGT occurred (i.e., the degree of difference in similarity of DNA between related and very unrelated species). “A” is the gold standard, with very high probability of HGT, while B and C have less assurance, but still probably still reflect HGT.

What Crisp et al. revealed was a moderate but not high frequency of HGT in two of the groups, and a low frequency in the other. The results suggest that there is indeed HGT, it’s not vanishingly rare, and that it can contribute to evolutionary change. But the level of HGT is not high enough to either efface phylogenetic trees or suggest that we revise evolutionary theory to say that genetic variation comes more often from HGT than from simple mutations in organisms.

The salient findings:

  • First, a refresher: primates have roughly 20,000 genes, fruit flies about 15,000, and the nematode Caenorhabditis about 11,000. You can see that there’s not much difference in gene number between these species: a result that surprised evolutionists and developmental biologists when the data first came out. Perhaps solipsistically, we humans seem a lot more complicated than flies, but don’t have many more genes. The difference may reside in how those genes are used, that is, in the regulation of a fairly constant number of genes.
  • How much HGT has occurred? In primates, the number of genes that have moved by HGT in classes A, B, and C are 32, 79, and 109, respectively. In Drosophila the numbers are 40, 25, and 4. In the nematodes it’s 68, 127, and 173, respectively.  There’s much less HGT in flies than in the other groups, but still the extent of HGT is only moderate in worms and primates: about 0.2%- 1.5%, depending on which species you use and what class of genes you want to count as acquired by HGT. That’s not enough transfer to constitute serious problems for making phylogenetic trees.
  • Most of the genes transferred to all three groups were those producing enzymes, which makes sense since they can confer immediate new functions on the recipient organism. Genes most often transferred affected the immune system, lipid metabolism, the modification of other large molecules, proteins produced when organisms are stressed, and antioxidant activities. This held across worms, flies, and primates.
  • Curiously, one gene that may have been transferred horizontally is the Landsteiner blood group gene in primates: the gene producing different antigens on red blood cells that give us type A, B, AB or O blood.
  • Finally, by placing the putatively transferred genes on the family tree of metazoans, they determined that gene transfer is both ancient and ongoing: HGT genes were acquired in both the ancient parts of a group’s phylogeny or in the more recent parts: say in the lineage of only one species of primate—which implies recent HGT since that one species diverged from other primates.

Where do the genes transferred into primates, worms, and flies come from? The authors provide this handy chart:

Screen Shot 2015-04-10 at 8.12.35 AM

The recipients are in the column to the left, the donor groups along the bottom. As you see, most genes come from microbes: bacteria and protists. This isn’t surprising because those organisms can transfer genes by either ingestion or infection.

The upshot is that we have a nice new finding, with some surprises—the ABO blood group genes still amaze me, and I’d like more confirmation—but a finding that hardly endangers evolutionary theory. And though I find the extent of HGT low, especially in flies, the authors do try to make a case that it’s rather high. As they note in the paper’s conclusions:

Although observed rates of acquisition of horizontally transferred genes in eukaryotes are generally lower than in prokaryotes, it appears that, far from being a rare occurrence, HGT has contributed to the evolution of many, perhaps all, animals and that the process is ongoing in most lineages. Between tens and hundreds of foreign genes are expressed in all the animals we surveyed, including humans. The majority of these genes are concerned with metabolism, suggesting that HGT contributes to biochemical diversification during animal evolution.

Well, technically “173” is “hundreds,” but I think the actual numbers I gave above, and the percentages of genes in a genome acquired by HGT, show that its prevalence is less than the paragraph above would suggest.

Nevertheless, it’s clear that we have a new source of genetic variation, and one that can contribute to adaptive evolution. But remember as well that the frequency of HGT transfer must be low, since such events must be rare, so we’re not entitled to say either that the number or commonality of HGT events is substantially higher than the number of genetic novelties produced by the more conventional process of mutations occurring within a species’ genome. This is why the branching bush of life seems, for the time, secure.


Crisp, A., C. Boschetti, M. Perry, A. Tunnacliffe, and G. Micklem. 2015. Expression of multiple horizontally acquired genes is a hallmark of both vertebrate and invertebrate genomes. Genome Biology 16:50, doi 10.1186/s13059-015-0607-3.

232 thoughts on “How often do genes move between distantly related species?

    1. Ditto! If ya’ll were to post on low-quality science articles, then I might have a bunch of comments. As it is, I tend to enjoy the post and tuck away its lessons in hopes of being able to draw upon it in some future conversation.

      I could have noted that just last weekend we saw an okapi in the Cincinnati Zoo, and that its fur looked invitingly velvety soft, but really, would anyone want to read this comment?

      1. I agree completely with Robert and Charles. I am a lifelong science nerd, but not a scientist. Since leaving fundamentalism, I have developed a special interest in evolution. I find it wonderfully fascinating!

        I came originally to this site for information on evolution, and I keep coming back for that and any other science. I never skip over a science article. However, by the time I read a post, there are usually 35 to 100 comments already; and I seldom feel I have anything unique to contribute.

    2. I second this.

      I’m not a scientist, but as a professional nerd, I’m familiar with the scientific method.

      I’m a retired infantryman who now works in the software industry as a tax analyst programmer. How Exciting (yawn)!

      As an atheist and military nerd, I feel comfortable commenting on the topics I know such as religion, history, politics, and anything military.

      But I do thoroughly enjoy all the science articles and read every one.

    3. Like many other readers, I don’t comment on science articles because I am not an expert on the subject and have nothing useful to add. But on the question of whether Hillary would be a good President – now that’s a subject I am an expert on.

      1. Like others I most like the biology posts. I come here for my education. For instance I learned something today in that, while I knew that horizontal gene transfer happened, I always thought it was a bacteria thing, I had no idea that it was so widespread a phenomenon.
        The wildlife photos are pretty and the noms make my mouth water but it’s only the science that changes me by giving me something new to think about.
        Thank you and keep it up.

    4. read with keen interest. I am much more likely to skim the general content and reserve the science posts for later slow and deliberate digestion but would rarely if ever comment (this one being the exception). I like the mix of content here on this website.

      1. Same here. I read most of them if I have time.

        But as a non-scientist, all I can really say in honesty is ‘that’s interesting’, which isn’t that interesting as a comment obviously!

        1. Same for me. I read them. I also read the Reader’s Wildlife Photographs posts but have never commented there.

      1. Also, One category – religion is opinion/fantasy/craziness, the other biology is fact/theory/knowledge.

        I think it is easier to comment on the first category, because everyone is as much an expert as anyone else (at least in their experiences and to the extent one can be an expert on a body of knowledge that is not knowledge)

        The second category is harder to comment on because well opinion versus facts again, unless you want people to just acknowledge that they have read the article, which I hope you can see lots of people do.

        I for one want to see these posts as much as any other.

        So here is a comment that is a question for Jerry or the authors or other experts:

        What about viruses transferring genes from mammals to mammals or even bird viruses transferring bird DNA to mammals etc? How much does that occur? Or mosquito or other vectors transferring DNA from one bitten animal to another of the same or different class?

        1. The two elements certainly occur. Some viruses integrate into the host genome as part of their life cycle and can take host genes with them when they replicate. Herpesviruses certainly seem to have picked up host genes. Other viruses – most notably flu, HIV, mouse mammary tumour virus etc – have definitely jumped species and/or can infect multiple species. Put these elements together with a sexually transmitted virus and I you have all the pieces in place for HGT. However, the more distantly related the hosts, the less likely this kind of cross-species infectivity would be. I’m not sure if it can happen between kingdoms. Something else, like bacteria, might be responsible for that. There are pathogenic bacteria with secretion systems that inject material into the host. Usually proteins but maybe DNA too in rare cases? (I don’t know.)

  1. Regular reader, first-time commenter. Just wanted to say that I come here for the science posts (and the cats and honorary cats). I don’t have much to add because it isn’t my field, but I appreciate the level of explanation provided, and the work involved. I often share posts with the bio students I tutor when they’re relevant to the topic we’re reviewing.

    1. Welcome! I save a ton of WEIT science bookmarks ‘just in case’. Your students are lucky to have this sort of exposure. You never know what a single experience will stir up in a young person.

  2. Actually, I only read the science posts, unless it’s REALLY interesting.

    But as a relative amateur, I don’t comment on them that much.

    1. Same here.
      When I come here, I first look for new science posts. I read the others only if I have the time.
      I cannot say that I always understand every detail, and often have to google to get at least a general idea about certain terms. However, with a very few exceptions, I find these posts very interesting and enriching.
      I almost never have anything of value to contribute, so I just shut up.
      I should also note that quite often, I also find great value in readers’ comments on these topics.

      1. I can also second that. I have noticed that Jerry’s style of writing is quite different on the science posts than on his religion posts. He reverts – totally appropriately to his base-state of writing — science writing. This style is very different than his religion post writing, which is easier to read.

        But this is not a criticism – it’s just an observation. Without sounding too elitist, I think this is why students in high school find science classes “harder” than religion classes or some of the “softer” social science classes (for example). The difficult or unfamiliar word density is necessarily higher in science education than most other subjects. This makes studying and reading more difficult — but also richer.

  3. Agree with Jay. The pure science articles are why I subscribe to this site. Your other content, while informative, is much more liable to pull in comments based on its nature. But I can get that type of content many other places on the web.

  4. Yes, Jay is right. Unless one is a trained scientist, and a specialist in the field that is being discussed, one is not really in a position to offer an opinion or to voice one’s agreement or disagreement. I certainly read and enjoy the scientific posts, but if there is no extraordinary error in argument that a layman can recognise, then there is not much that one can contribute in the way of comment. So please keep up the scientific posts!

  5. The percentage of subscribers able to meaningfully expand or correct your science posts is far smaller than those willing to opine on other topics. Like Jay, I very much enjoy these posts because I’m learning something. So, a comment here seems more like an intellectual selfie, rather than something more substantive.

    Or, perhaps we should all recognize that an attaboy isn’t necessarily the same as an intellectual selfie.

    Thanks for keeping a portion of my day focused on thinking and learning something I don’t know!

  6. As a non-scientist, I must (sort of like Maru) say that I do read the science posts, even though I usually understand only the broadest outline. I kind of enjoy the challenge, but like jay above, there’s not a lot I can add to any discussion. Sigh.

    1. That’s pretty much me too. Always read them, almost never comment as I’m completely incapable of adding anything intelligent on such posts. I did comment on Greg’s post the other day, but it’s not a comment that would have stood out for its erudition.

  7. I have never posted here before, but have been a regular reader for many years.

    This time I just wanted to say that I did read the post on Okapi, and found it very interesting. Reading about evolution and natural history is what made me ditch faith and become an atheist. (Dawkins’ God Delusion was the final nail in the coffin, to be fair)

    I appreciate the work you put into the postings here, and educate myself on these almost every day. Thank you!

  8. I imagine part of the reason, at least, that comments for science posts are less than for others, is that we lay readers cannot say too much that is salient on topics of detailed science, other than, perhaps, “wow” or “very interesting”, but on religious/philosophical/social topics we can all manage an opinion, if we choose to express it. That said, this is a very interesting topic which shows the interrelated nature of living things, and the complexity of evolutionary processes. Wow. On a science topic other than this, but of an evolutionary nature, have you a comment on the recent study of the heights of the Dutch, suggesting the recent increase is driven by natural selection? I am skeptical; I think the time period is too short to illustrate natural selection. [Does natural selection favour taller stature among the tallest people on earth?

  9. I think the last point you make is important and deserves to be underlined. The number of genes in a given species that have horizontal transfer some time in their past is not a good proxy for the rate of horizontal transfer, just as the cumulative number of any event is not a proxy for the instantaneous rate of such events.

    There’s been a similar misunderstanding in botany that resulted in wildly high estimates of the rate of hybrid speciation. A large percentage of angiosperms have a hybrid species somewhere in their ancestry, but hybrid speciations are a small minority of speciations. Very rare events add up over time.

    1. John hits the nail on the head (as he usually does). A lot of people hear how many horizontal gene transfer events have occurred in the ancestry of a species, and mistakenly think that this means that HGT occurs during most reproduction events.

      The extent of HGT observed means that the overwhelming majority of new offspring do not contain even a single gene newly acquired by HGT. Transmission is overwhelmingly vertical rather than horizontal. It’s just that there have been a lot of generations for it to occur.

      1. “The extent of HGT observed means that the overwhelming majority of new offspring do not contain even a single gene newly acquired by HGT.”

        Joe, I think that should be qualified; the rate of entry of horizontally-transferred genes per individual might be higher than the study suggests, since most genes acquired by HGT will not get fixed.

          1. I think my only piece to say on this subject fits here. When I was an even more newbie on genetics, as interested in astrobiology I early on met the claim that HGT could (should, would) be prominent and possibly make deep trees nonsensical. The problem being compounded by the regular use of just 16S rRNA to place prokaryotes.

            When I later saw (early and naive, of course) attempts to estimate the HGT fixation rate ending up ~ 1 HGT/gene I became less worried. One HGT fixation/gene over a few billion years seems naively tractable. I hope I am not wrong?

  10. I read a book about this: The Mystery of Metamorphosis: A Scientific Detective Story by Frank Ryan. His hypothesis is that the metamorphosis seen in insects and sea life is due to the fusing of two distinct species (and their genomes). A wild idea, supported by Lynn Margulis, who wrote the forward for the book. The book itself has been heavily criticized in the biology community; I found it a very interesting and well-written book, even though its central idea is quite farfetched.

      1. Thanks for the link to the review. I did find it entertaining. I don’t have a copy of the book (I got it from the local public library), but I remember being impressed by Ryan’s reference to metamorphosis in a species of squid as evidence for his hybridization hypothesis. Of course, this is one species of many, but nevertheless…In any case, the book is a polemic supporting an idea, and not a balanced view, as your review explains.

    1. It should be pointed out that this theory would be easily tested by looking into the genomes of the organisms involved to see if there’s a massive signature of horizontal transfer. It was, and there wasn’t.

    2. I don’t buy it. The genes used to develop an embryo into a larva are for the most part the same genes used to transform the larva into an adult. They are mainly expressed in different spatial patterns and turned on in different tissues to remodel the larva–> an adult.

    3. When you have a really good but at the time seemingly wild theory (like L.M. does/did), then everything shaped like a nail looks like it can be fixed with a hammer.

  11. I also read the science posts but refrain from commenting because I’ve nothing new to add except lame little “oh wow”s or such. Sometimes I have a question but I’m not sure if it’s a good question — or one which will go off on a tangent (or reveal that I couldn’t possibly have understood the issue.) For example:

    For the transfer of genes between distantly related species is simply a new source of genetic variation—like a mutation—whose fate is still subject to whether the transferred genes are good or bad for the host.

    What if the new genetic material is neither good nor bad? How many genes (and characteristics) which are pointless from the standpoint of survival can accumulate before it eventually becomes too costly for the host? I’m not sure what would set that limit, or how we could test it.

    I’m imaging organisms which look like something out of Dr. Seuss.

    1. Yes, Sastra, that’s a good observation. Jerry was simplifying when he said “In the case of aphids, for example, the acquired pigment genes rose in frequency in the pea aphid species by natural selection, but had they been deleterious they would have been eliminated.”

      If he were writing a scientific paper, he probably would have mentioned that neutral or nearly-neutral HGT is the most likely kind, and these have a certain probability to go to fixation, with the probability depending strongly on the population size, just like any other nearly neutral mutation.

      I think Joe’s and John’s observations in Comment 14 need a small qualifier in light of the possibility of gene transfer of neutral genes. The number of horizontally-transferred genes that enter an organism per generation may be slightly higher than their comments suggest, because we are only seeing the ones that get fixed (and at least in primates with their small population sizes, these are probably fixed due to natural selection). Neutral and deleterious transferred genes would be eliminated by drift or negative selection. But they may still be entering genomes at a decent rate per individual per generation. They just aren’t sticking around much. As far as I can see, this study doesn’t tell us much about the actual rate of HGT per individual.

    2. Your question is a very good one, and a current controversy in biology.

      It’s hard for an organism to get rid of unnecessary DNA without loosing the good stuff too. Therefore, useless stuff accumulates. Some of the useless stuff can come from HGT, some from mutated genes of its own that no longer work (pseudogenes), from sloppy copying that creates extra copies of DNA, from RNA copied back into the DNA (e.g. Alu in humans), from viruses and transposons. This useless stuff is often called junk DNA.

      How much junk DNA can be tolerated? In most species, a lot. Between 65% and 90% of the human genome seems to be junk. Some plants seem to have much more.

      Organisms under severe selection for biochemical efficiency seem to loose the junk DNA sometimes. Many bacteria have really lean genomes. The bladderwort Utricularia (a plant) lost most of its junk DNA. Apparently the fugu, a pufferfish, did too.

      1. Interesting points about the likely secondary loss of junk DNA in some lineages. I am not sure why there would be selective pressure to loose junk in pufferfish, though. Pressure that I can think of would be for smaller nuclei (as in lightweight birds), or shorter cell cycle times. But why would a fish need these things?

  12. I too read the science posts, but am less likely to comment. Thank you for continuing to make the effort.

  13. Implied in Jerry’s post, but I’ll emphasize, is the stunning difference between prokaryotes and eukaryotes on the HGT (also called lateral gene transfer or LGT) axis. Prokaryotes are basically one gynormous gene pool, flipping genes between themselves. A paper with cool graphics on this point is Dagan T, Artzy-Randrup Y, Martin W: Modular networks and cumulative impact of lateral transfer in prokaryote genome evolution. Proc Natl Acad Sci USA 105:10039–10044 (2008).

  14. One thing, as a suggestion, would be to delineate the science and non-science posts somehow.

    I didn’t realize that the Opaki post was a science post until you mentioned it. I thought it might have been just more pictures, which I generally skip.

      1. Also maybe you could (I don’t know how) have a “like button” or something like that with the science posts, just so that you could have some feedback/assurance that these science posts are indeed being read.

        It certainly seems like the posts are being read, but most readers might feel silly taking up comment space saying “cool” or “wow”, but would probably click a button to let you know the article has been read.

        1. WordPress does have a “Like” button already. Not sure if you need to be logged in to click it. My “problem” is that I generally read the posts in my email, often on my phone, so I can’t always click the “Like” button – and probably don’t even count as a page view. If Jerry wants to avoid this, I think you can get the email to only include the first few lines, so people have to visit the page. (Email alerts from coelsblog are like this.)

  15. I want to echo the others about commenting on the scientific posts. I read them but don’t often comment.
    That said; I could see anti-GMO fear mongers latching on to this.

    1. I think it’d work the other way: See, transgenic organisms are Natural! And as any anti-GMO fear monger knows, Natural = Good.

  16. I also read mostly the science posts. Without them, I ink I would leave. Keep them coming.

    It is not surprising that the frequency of HGT is low in complex eukaryotes. The only way for them to be retained in a species is for the genes to be transferred to the germ line in the gonads. At least in humans, infection of the gonads is very uncommon and can result in sterility.

    It would be fascinating for someone to develop a vertebrate model that shows exactly how this can occur.

      1. If entry is in fish and amphibians then it should be easier, at least in species that develop as exposed eggs in a wet environment.

      2. That is my question too. Maybe there are answers farther down in the comments. Could the genes (or segments of DNA) somehow be present during meiosis and get stuck into the chromatid (if that is the right word) much like crossing-over happens?

        1. I have been thinking about this all day with regard to vertebrates. Is it possible that there are microorganisms that can live in vertebrate gonads without doing any harm? I am thinking primarily of the testes, which are an immunologically privileged site. The ovaries are not.

          This would provide a potential mechanism for horizontal gene transfer. In the past, I would have rejected this idea summarily, and I think most physicians and scientists still would, but given the evidence for HGT, I have to wonder. I do not think anyone has looked for this.

          1. One highly hypothetical way would be during embryonic development where the early germ line cells are small in #. Any entry into the germ line would then lead to a high # of descendant germ line cells having an insertion.

          2. But, still…how does foreign DNA make its way to the nuclei of germ-line cells of an embryo in a form such that it gets incorporated?

            We’ve got a smoking gun of an whodunnit but still, best I can tell, absolutely no clue about how it was actually done. Seems to me like there oughta be people lining up to write grant proposals to figure this one out — especially considering how lucrative genetic engineering is. Maybe the mechanism can be replicated in the lab in a way that would be more efficient than anything we’ve currently got so far?


          3. The one word answer might be: viruses.

            Transfer of genes from one organism to another can be carried out by viruses, and they can get to germ line cells too.

          4. Yes…but from plants to primates?

            I wasn’t aware that there are any viruses that can infect both…which means a really long chain of hosts and viruses to get the DNA from the one to the other, which would, in turn, leave its own evidence trail.


          5. As these are extremely rare events, I guess there could be more complex courses of events involved. e.g. A macrophage eats a bacterium and is also infected with a virus that picks up some of the bacterial DNA. Insect vectors and multispecies parasites like malaria could potentially act as shuttles for odd cross-species transmissions without a virus that can directly infect the two taxa that we now detect those genes in.

            Personally, I think that the majority of cross-kingdom HGT that this paper proposes are more likely to be methodological artefacts, though. (They make no estimate of false positive rates as far as I can tell, and there are lots of things that can go wrong with that kind of inference.) In which case, these are REALLY rare events.

          6. Thanks for that. I’ll remain skeptical, especially about the most distantly-related examples (such as plants to primates).

            I could see transfers between birds and primates, since influenza sweeps between the two populations in such large numbers…but, even then, I’d expect it to be vanishingly rare for any such transfers to make it into the germline. Anything beyond that…well, I think I’d want to see some mathematical modeling including a proposed mechanism and rates of occurrence and expected outcomes before having much confidence in the lab results. Or, conversely, seeing multiple labs reproduce the same results with different methods — that would be rather convincing, too.


          7. I think skepticism is definitely the way to go! It’s easy enough (in principle) to find something that looks like a smoking gun (i.e. DNA sequence without close homology to any known DNA of the right taxonomic origin) but very hard to then be sure that’s what you have. With our virus project, for example, we might find some sequence of viral origins in a genome (and I’ve written a program to do this) but the really hard bit is going to be ascertaining which virus it came from. Just because the closest known virus is a plant virus, that doesn’t rule out a related virus with a different host range that we simply don’t know about. (I reckon there are billions of different viruses out there, most completely unknown.)

            To be fully convinced, I don’t think a smoking gun will be enough – we’ll need to catch one in the act. Not impossible but very difficult. In the meantime… well, we all know that extraordinary claims need extraordinary evidence!

          8. Cabbages of Doom,
            This is great.I am getting a whole education.thanks.
            I have an excellent background in genetics and whatever but am wildly out of the loop.
            And thanks to Prof. CC for providing the opportunity.

          9. Thanks for the clarifications…and I hope to see Jerry posting the results of your research sooner rather than later!

            In fact…maybe do a write-up of the project for him to do as a guest post in the mean time? I have a sneaking suspicion it might be at least as popular as this 200-plus-comment post, especially if you’re active answering everybody’s questions.


  17. Another lay reader who likes the science articles but seldom comments.
    Given the number of busted virus genomes and such in our chromosomes, it is not surprising that other bits of DNA sometimes get incorporated (perhaps with the help of viral enzymes designed to carry out that exact task?) Still, to be inherited, the incorporation would have to be in germ line cells or fertilized zygote, and do something useful to be selected for. That would imply that the novel DNA be transcribed and do something immediately useful, or be maintained or increased in the population by genetic drift until mutation or circumstances make it useful. So it’s not surprising to me at least, that the process is fairly rare.

  18. I can understand why the lack of comments on science posts is dispiriting, given the work that goes into them. I read the Okapi post and found it very interesting, but had nothing to add. I usually just skim the genetics posts because they’re over my head. Your writing style is friendly to skimming, and I mean that as a compliment.

  19. “The phenomenon of HGT has been called “non-Darwinian,” since it simply wasn’t envisioned by Darwin or, indeed, even in the early days of the “modern synthetic theory” of evolution. But, contra some critics of evolution, HGT does not invalidate the modern theory of evolution.”

    This is an important point because science writers often get into the lazy habit of casting each new discovery in evolutionary biology in terms of “what Darwin knew”(“Darwin was wrong!”). This habit is absurd. Even if the frequency of HGT were higher, that would have little to do with the state of scientific knowledge in Victorian times! Think about our current understanding of genetics (apparently Mendel was “wrong”!). It would be more accurate to call HGT, non-Mendelian than non-Darwinian, but both adjectives are completely unnecessary.

    Darwin was indeed wrong on many things – he thought, for example, that the Cambrian period was about 65 million years ago, and his understanding of inheritance was minuscule by today’s standards – but how could it be otherwise? A bright student in my INTRO biology class could give Darwin a tutorial on the particular mechanisms and consequences of evolution.

    I think we biologists are largely to blame for this because of the still-frequent use of words and phrases such as “Darwinism” (as a synonym for modern evolutionary biology, which it isn’t), or “Darwinian natural selection” (do we really need the modifier to distinguish it from “Wallacian natural selection”?). “Darwinism” is particularly odious because it feeds into the notion, promulgated by creationists, that evolutionary biology is an ideology. No one should downplay the significant contributions of Darwin to biology, but this constant reference back to Darwin is largely counter-productive for explaining biological phenomena.

    1. I just remembered: Yet another lazy habit of science writers is to greet EACH new discovery in human evolution (and they are coming in fast and furiously these days) as “shaking the family tree”. PUH-lease.

    2. I think HGT shows that Darwin really was wrong on one important point. He was sure that evolution proceeded by very small steps. HGT and a few other processes show that this was not exactly correct. A new trait can be acquired suddenly and fully-formed, though of course this trait did evolve in small steps somewhere else.

        1. I think that would depend on if you are thinking of an organism-centered or gene-centered view. (“More likely yes” respectively “more likely no”, at a guess.)

      1. Darwin did allow that the rate of evolution might speed up or slow down on occasion, noting for example that the modern brachiopod Lingula is not much different from Silurian times.
        In a later edition of the Origin he seemed to toy with a thought that seemed similar to punctuated equilibrium when he wrote: “the periods during which species have undergone modification, though long as measured in years, have probably been short in comparison with the periods during which they retain the same form.”

    3. Despite what the detractors say, I think that Darwin gets much credit with coming up with a remarkably, although not entirely, complete theory without knowing what DNA or a gene are and without any understanding of heredity. His achievement is right up there with the leaps of some amazing geniuses, including Einstein.

    4. Agree completely. Along the same lines the typical “new finding proves modern evolutionary theory wrong” is equally tiresome. It demonstrates ignorance of the very basics of how modern science evolves Theories as well as the specifics of the primary aspects of the modern Theory of Evolution.

      What is really a shame is that some biologists play this card too. They should know better. And I mean that ethically and otherwise, as the specific case may be.

    5. I need to qualify this comment by stating that I’m talking without authoritative expertise. None the less, I think it’s correct.

      The Darwinian modifier is due to the christians trying to modify evolution theory to include their magic pixy dust. The Darwin modifier isn’t about distinguishing between different legitimately founded process proposals. Due to the overwhelming evidence that evolution is the basis for variation in life forms, the christians want to claim that evolution is guided by some godish thingy and legitimate scientist have needed some way of stating that there are no gods involved.

  20. When referring to a gene, that has moved by HGT, does this mean a sequence of bases that we know represent a particular gene? Not just a sequence of bases that probably represent a gene?

    Furthermore, does it mean that this gene is definitely active in the new organism, or were they only checking for presence?

    Under natural selection we would expect that only active and useful genes would become fixed in the target population?

    I’m just wondering if this sort of thing can be worked into another of Jerry’s tests for evolution, with natural selection predicting the presence and activity of genes acquired by HGT to follow a certain pattern.

    1. That’s a good point. The study specifically focused on expressed genes (messenger RNA transcripts) as a way to test the functionality of sequences that were also found in the genomic sequence (DNA). Seems like reliable confirmation that the transferred sequences are functional.

    2. “Under natural selection we would expect that only active and useful genes would become fixed in the target population?”

      Not always. It doesn’t have to be useful to become fixed. Even mildly deleterious things can become fixed in the right circumstances (e.g. small population sizes).

      It is an interesting question, not addressed by this paper I don’t think, whether or not these genes are actively expressed and yield active protein products. My expectation would be that some are, some aren’t, and that the ones that are expressed are more likely to have been fixed by selection and the ones that aren’t are more likely to have been fixed by chance.

  21. Please keep the science articles coming. I, as is also reflected by other commenters, do read them with gusto. They keep my mind streching and broaden my perspective in many areas. I have always been a generalist and love reading and studying new ideas or new ways of looking at old ones. At 74 yrs of age WEIT is always a bright spot each day.
    Keep up the great work you are doing. Presently I am impatiently awaiting the Albatross. Thank you for enriching my daily life.

    1. My life has also been enriched by the science articles and all the rest of the content provided in WEIT. By all means keep it all coming.

  22. Is there any evidence that HGT decreases in lineages as the get older since there might not be as much scope for novelty? For example say the metabolic machinery of an organism is already well developed / settled, and there’s no selection pressure wherein HGT of a novel gene would contribute an advantage – would that constrain HGT to early development of a lineage?

  23. It was a long time ago and my memory is a bit hazy but, when I was an undergraduate, Lynn Margulis’ ideas about the origin of cellular organelles were considered controversial. I think she was prepared to attribute the origin of just about any identifiable organelle to an originally independent organism (at least that’s how her ideas were presented). However, the idea seemed almost incontrovertible with regards to mitochondria and chloroplasts.

    Looking at this new information, it occurs to me that the transferred genes are not going to be inherited if they are not incorporated in germ cells. Not knowing the mechanism, may one suppose that by chance this will happen more frequently in somatic cells and, therefore, might it be the case that any individual organism might carry more HGT genes in its somatic cells?

  24. Nice summary of Crisp et al.

    Margulis was not the first to suggest a bacterial origin of mitochondria; the idea is about 100 years old. She does get credit for the modern formulation of the hypothesis,

  25. I think it’s an interesting but alarming idea that eating an organism which is engineered to make pesticides could lead to the consumer, or the bacteria in the consumer’s gut, to gain those genes and to start generating those pesticides in the gut.

    Can that idea be ruled out? I don’t think one could do so, because that’s the kind of reasoning some use to argue against evolution. “The odds are too small! It just can’t happen!”

    It’s an interesting thought, and a possible talking point against GMO foods.

    1. Aside from the extremely low probability of the genetically modified gene (of all possible genes) from GMO food (of all possible foods) would get incorporated into the human genome, most (possibly all?) genes used to modify our food organisms occur naturally in other species. The genes come from bacteria or diverse species of plants. So if you’re going to worry about getting these genes from GMO’s, you should also worry about getting them from your spinach or broccoli (actual sources of certain genes being transferred into other crops now — in researach, not yet release as far as I know).

      There are many important, urgent issues we need to deal with to protect our environment and ourselves. GMO foods just aren’t on the list, at least for me.

    2. Ingested raw food is more likely to transfer genes to your bacterial symbionts. For example, many Japanese has a gut bacteria that got its algae eating enzymes from a marine microbe. [ ]

      But generally targeted artificial selection (such as ‘GMO’) is more controlled than the broad selective sweeps our old methods have used for tens of thousands of years. Those bring slightly harmful genes with.

      I would worry more about fully uncontrolled infections though, ~ 7 % of our genome is viral heritage.

      Many viral genes are very useful though, the immune system regulation genes that allow a tightly embedded placenta despite a fetus nearby are viral. That hominid placentas are well supplied with nutrients and allow a long pregnancy is likely one of the reasons why we can grow a large brain.

    3. HGT exists, as is the point of this posting, but it is very inefficient.
      Even if a gene for making an insecticide were transferred to some gut cells, the effect could not possibly be of concern b/c the dosage would be far too small. I know of no such GMO that would be of concern to us even if the dose was large. The gene for BT endotoxin, which is an insecticide used in GM crop plants, has no effect on humans. It is highly specific to only certain kinds of insects.

    4. The mechanisms of HGT are apparently not well understood. I briefly perused the literature and ran across much handwaving and little to clarify the question that you ask. It’s unclear if eating, say, GMO rice that has genes to produce flavenoids, has any liklihodd of transferring that synthetic ability to your intestinal epithelium. Even if some cells incorporated the genes for the synthetic pathway, they’d quickly be shed and replaced with cells that don’t have the genes. Similarly, there is no obvious way that the genes in intestinal epithelium could make their way to other cells in the body. Most articles seem to favor HGT in higher eukaryotes occurring via viral infection. This makes a certain amount of sense, as the virus could infect germ cells and potentially insert genes that could be inherited. But this mechanism is lacking in the case of simply ingested foods.

      1. That was going to be my question: how do ingested genes get from the gut to the germ line? The only vaguely plausible mechanism I could think of is for genes ingested by a pregnant mother to be somehow absorbed by a developing embryo before significant tissue differentiation.

  26. I very much enjoy the science posts and read them thoroughly. They anchor all the other delightful and surprising stuff that shows up.
    I have a question about HGT, maybe naive–how does a gene transferred into a eukaryote, presumably from the periphery, get into the germ line? Just happenstance?

    1. I would be absolutely delighted if somebody would answer this question.

      Especially…how the fuck do plant genes make their way into primate germ lines!?


    2. Prokaryotes are quite happy to swap genes [ ] or, I think, express RNA and DNA that they ingest.

      Moving on to eukaryotes, they would early on be similar after the split from archaea (unless that was the reason for the split, which is unlikely). Maybe they still did so when the pre-mitochondria started to parasitize them. (Which the latest, most comprehensive mitochondrial phylogeny says.) There is a hypothesis that the nucleus evolved in order to protect the core genome from too much transfer.

      But different species have different number of core genes that derive from mitochondria. Perhaps they have lost genes, But perhaps a number of genes were transferred from mitochondria after the nucleus evolved.

      I have very vague ideas of how that would happen, but I have seen people mention that dead mitochondria (which numbers several hundreds in eukaryotes) may at times have their genes incorporated during eukaryote cell divisions, from mitochondria lysis (or however dead mitochondria are taken care of) and eukaryote repair mechanisms at a guess.

      If it happens in the germ line, the HGT genes are inherited, and may be expressed and sometimes fixate. The mitochondria will see them as duplicated if they can import their products, and if so they may lose their own copy. There should be a selective pressure to loose endosymbiont genes.

      I would have to hunt up the reference, but there is a very interesting endosymbiont system in some insect organs. (They inherit the symbionts into special nodules with the maternal egg, IIRC.) An insect endosymbiont bacteria has, in some insect species, its own endosymbiont in a matroshka fashion. The insect but also the innermost symbiont have gained genes by HGT, IIRC.

  27. HGT does not invalidate the modern theory of evolution. For the transfer of genes between distantly related species is simply a new source of genetic variation—like a mutation—whose fate is still subject to whether the transferred genes are good or bad for the host.

    I think its seen as somewhat non-Darwinian because while it is ‘descent with modification,’ its not ‘modification from descent’ like sexual gene mixing. Of course, many other mutations also don’t ‘come from’ either parent too, so HGTs should not really be treated as any different from, say, a radiation-induced SNP.

    1. But evolution is gene centered. (Dawkins’s selfish gene.)

      The gene think it has descended as per usual. Juts by a funny mechanism, and into a new type of cellular vehicle.

  28. This is the reason why I gave up on my major concern on genetic modified organisms. I’ve never been concerned about eating food GMOs, but I was concerned about the safety of gene transfer between species that never experienced gene transfer. I think it was someone here who pointed out this already happens.

    Why would evolutionists be concerned?
    Because creationists will misuse anything and everything to to try to discredit evolution? They do that anyways, they are ignorant of evolution, or dishonest or both.

  29. I did read the Okapi post yesterday, or rather scanned it, and it was fascinating on many levels. The problem with the science posts is that they not only take longer to write, they take me, at least, much longer to read. I not only have to understand what the writer is saying, I need to orient to the context in which he/she is saying it. If I’m in my own academic territory, that’s fine, but if I’m not, it takes a bit of doing.

    My solution is to flag the article for reading later when I have more time. (The Okapi post is flagged.) I will probably not really read this article or the Okapi post until next Sunday. And it doesn’t feel right to post a comment after the conversation has moved miles down the pike. Maybe I’m wrong about that. Anyway, the science articles here are terrific and the heart of the site. Keep them coming!!

  30. It would be interesting to know how often the same taxa show horizontal transfer of non-coding sequences. Would this help to test the controversial assertions by the ENCODE consortium that most of the genome is functional and under selection? One possible prediction: a much higher rate of HGT for noncoding sequences (what ENCODE enthusiasts think of as regulatory sequences) would argue against their functional importance if animal genomes can tolerate a much higher rate of insertion of those foreign sequences into their chromosomes. The paper looked at introns (non-coding sequences within genes) but not (so far as I can tell) at regulatory regions (noncoding sequences between genes).

    1. Obligatory correction: “non-coding” doesn’t mean “non-functional” or even “claimed to be non-functional”. You mean the last. Of course the great majority of non-coding sequences are considered junk, but we should watch our language. For one thing, every time a non-coding sequence is found to have a function, it’s taken as proof that all non-coding sequences have functions, even if that particular bit had been known to be functional for decades.

      There’s also a bit of a discernment bias in your proposed test. Functional sequences tend to be conserved by selection, but non-functional ones change freely, which means the detectability of such transfer events degrades much more quickly over evolutionary time. After a while, the functional insertions are the only ones you can see.

      1. Yes, of course, noncoding is not the same as nonfunctional. That was equivalence was the naive view that ENCODE used as a straw man to propose that most noncoding sequences are associated with some important and potentially conserved function. The ENCODE enthusiasts make the strong claim that in fact most noncoding sequences should not be considered to be junk. Yes, for sure it’s a biased test. Do you think it would tend to make the rejection of the ENCODE proposition more convincing?

        1. I don’t actually think we need any more convincing. Would another monkey genome make the proposition that humans are primates more convincing?

          But it would certainly be a conservative test.

          1. Hah, yes, like you I am already convinced as well (that the ENCODE assertions are deeply flawed). Dan Graur is fun to read on that score. My understanding is that some others are not so convinced. But maybe this kind of HGT comparison among types of sequences would not convince them either.

          2. Wasn’t the main problem that when the ENCODE described “function” it was a certain level of RNA expression (let us call that chemical function) irrespective of cellular use (which would be evolutionary function)?

          3. Yes, expression plus some predictions about how an RNA transcript from a particular noncoding part of the genome would be expected to bind with DNA or other RNA molecules as promoters or repressors. The predictions were based on structural models of RNA folding and RNA-DNA binding. I don’t know how those structural model predictions work so I can’t say how good the functional predictions are, but others said that the predictions were poor. That was a big part of the basis for criticizing the ENCODE interpretation that vast noncoding parts of the genome are ‘functional’. The alternative explanation was that, basically, for many intergenic regions of the genome the DNA may be transcribed but those transcripts do nothing, get degraded, and so are not functional. At least that was my non-expert understanding of the criticism of ENCODE.

  31. Cool study. Aren’t viruses also a major source of lateral gene transfer? Granted, viruses generally tend to have very small genomes that code for a very small number of proteins but they also hijack and transfer host proteins that enhance their replication and virulence.

    1. I have a variant question: How often are virus genes useful? Are they always junk, or do some end up giving metazoa or prokaryotes new capabilities? I’d think the virus genes would have fewer useful functions for us to adopt, but what do I know?

      1. Virus genes can be useful. There’s a syncitin gene in some mammals (humans have it) which we inherited from viruses and is now essential for the formation of the placenta.

    2. Our genome is littered with endogenous retrovirus sequences. I don’t think this study includes them because inserting DNA into the host genome is their modus operandi.

      There are some non-retroviruses that end up in the genome as well, but those are much rarer events – those still more common than those covered by this study. Why they don’t have a virus column for the non-retrovirus HGT cases is a good question. Perhaps because all viruses operate by injecting their genome into host cells, which makes HGT vastly more likely, so that including viruses would make the unlikely non-viral cases of HGT look like background noise.

      1. So, we have viruses inserting their genomes into those of other organisms.

        How often does it go the other way? How often does non-viral DNA wind up inside a virus’s genome? And by what mechanism…?


        1. I couldn’t guess how often, but it’s known to happen. I don’t what the mechanism would be in any kind of detail, though I’d bet it involves reverse transcription grabbing host RNA as well as viral RNA when constructing new DNA for the viral offspring, which then carry that chunk of host-sourced DNA off to another host where it becomes replicated inside infected cells and gets the opportunity to find its way into the nucleus of the new host’s germ line cells. All staggeringly improbable, but inevitable given staggering amounts of time.

          I’d bet Abby Smith would have a much more satisfying (not to mention accurate) answer.

          1. I’ve actually got a project ongoing in collaboration with virology lab to try and answer this question to some extent, i.e. we are looking for evidence of viral sequences in genomes that should not be their normal hosts. I’ll try to report back if we find anything! (I suspect it’s extremely rare.)

          2. I’ll hold you to it! Even if you don’t find anything, I’d hope you’d at least write up that fact for Jerry to publish (even if the journals don’t want to publish such “didn’t find anything” results).


  32. Reading the science posts, such as this one do take more concentrated effort for those of us not in the field but that is okay because we still learn and that is why we are here.

    So this very layman type conclusion here is that this study shows HGT is relatively small, therefore, does not cause a problem with proper location in the tree. That it is responsible for the origin of mitochondria is another one I did not know.

  33. What about the effect size and rate of these events? Even though HGT is rare, shouldn’t the potential effect of such events be very large and rapid? Mutation is akin to pulling a book off your library shelf and changing a word or two before replacing it. Change will undoubtedly come, but most will not be useful. HGT is akin to getting new shelves of books in your library. If you can integrate the new material into the functioning of your library, dramatic change could occur immediately.

    1. My thoughts exactly. As a percentage that is a low number, but I had thought the number would be much higher for single-cell organisms and fungi, for example, and “vanishingly rare” for primates. very informative.

        1. There’s dust, too? ;o)

          In defense of my limited late 1970’s biology education, Wikipedia tells me HGT as a concept had been around only 30-odd years, and was not a confirmed phenomenon until after I’d taken my last bio class.

  34. Just to add my two pennyworth to the many other readers who read, enjoy and learn from the science posts. As a (very) ex-chemist I am hesitant to comment or pose a doubtlessly naive question for fear of coming across as a lame-brain. I should add that the rest of us are fortunate in those of Jerry’s readers who do have something to add to the discussion: I try to learn from them too, and appreciate the efforts they themselves make.

  35. In my defense, I’m in an airport, have been traveling, and haven’t been able to keep up. But I read ’em!

  36. I confess that I did not read the okapi story. I am generally uninterested in ungulates, unless they are turning into whales, and so I hardly gave the article a glance. I did skim the article headings to see if it had an evolutionary story or some interesting molecular biology, but it didn’t seem so. The migration and lifestyle of these animals does not interest me, so I moved on.

    I did read the earlier article about the weird fly, and spent some time studying the linked fly tree. Probably there is someone with the opposite preference. In any case, I find that only about one in three of the articles of any kind interest me. That seems like a pretty good ratio in my view, though. 90% of what I see in most print sources doesn’t interest me.

    1. BTW: The present article I read with great interest and care. HGT is a very interesting topic to me, both because I’m interested in the molecular details and also because of it’s implications for evolution.

      If you hadn’t called us out for not commenting, though, I doubt I would have said anything about it. I’m probably almost competent to talk about HGT, so unlike many here I won’t claim ignorance of the topic is the reason. This result is very interesting, but it doesn’t seem terribly surprising, or controversial to me(though fake controversy is sure to follow it). Absent someone trying to forcefully argue that it IS surprising or controversial, I don’t have anything to add.

      I think that’s the real crux of the low comments on science posts. I often visit Larry Moran’s website as well and the science posts there seem to get many more comments than the science posts here. I think that is because the science posts there are quite often presented in the context of an ongoing argument between ID proponents and the rest of science, with ID proponents actually showing up to argue their side in the comments. Science presented as a point in a debate tends to engender more comments in the form of debate.

      I’m in no way suggesting that the science posts here should be changed to create debate. I think they are fine as they are. I’m just observing that I think that the framing of science posts plays into how many comments they get. Just as asking readers what *they* would do on an airplane if an ultra-Orthodox Jew asked them to give up their seat so that they didn’t have to sit by a woman probably greatly expanded the number of comments that article got.

  37. I read pretty much all that is posted, then I go look up more information for those bits that I find interesting…


  38. I managed to go through my schooling without ever taking biology – in Canada, no less but due entirely to a scheduling bug when computers were first used in enrollment – and have been trying to fill in my vast ignorance ever since.

    I consistently read these posts and find them fascinating. But I don’t comment because the word ‘fascinating’ isn’t the least bit interesting for anyone else to read.

    So, please, continue doing these fascinating posts. You are educating people in the biological marvels of the world even if the comments are fewer. Speaking for myself, my level of appreciation for your posts is inversely proportional to the number of comments!

  39. Just a friendly comment about the value of the science posts, as opposed to the antitheist, cat, boot, nom, wildlife posts you do. Lots of us come here for lots of different reasons. I for one don’t care anything at all for boots, but as you correctly tell trolls, it’s your website! I myself love your antitheist and cat posts the most–and those amazing wildlife photos (and info!)–but I do read the science posts on occasion. Some of it is a little over my head as a philosopher, and most times I don’t really care about the details to the level some other people do. BUT, these posts are EXTREMELY IMPORTANT and valuable to lots of people. When I do read them, I rarely comment, but find them almost always well-written and insightful.

    Here’s a question, you also post some pretty deep posts on free will–do those posts also get fewer responses? On the order of the amount the science posts get? I predict they do. If I more or less right on this, we can say that as the technical content about a specific subject increases, the number of people reading and responding will go down. Naturally.

    But I support you in maintaining BOTH kinds (or even all kinds) of posts. A) You have absolute right to do it, B) some people love each of the palette of different kinds of posts, and C) there is no C.

    1. How can we be sure there is no C)? Does it exist in the abstract sense, or is it only called forth by a sense of will?

      (Ha, be nice to me. I am a science student, not philosophy)

    2. You will be happy to learn that Jerry’s posts on free will receive more comments than just about any other topic.

      I you do find such discussions interesting I highly recommend you search for free will articles on WEIT and read the articles and the comments. Lots of reading though.

    3. The free will posts also generate not only lots of comments but also, by my casual estimation, the most controversy and back-and-forth among commenters. And that is saying something: there is also a wide range of opinion expressed in posts that touch on the Israeli-Palestinian conflict, which, given the timeliness and seriousness of the issue, one might think would result in more debate than the relatively academic subject of free will. But no! Free will is a free-for-all!

      My theory is that free will is a relatively safe subject to debate, in that there is almost no risk of inadvertently making a comment that could be offensive or demeaning to a non-European cultural group. And also the Middle East is a difficult problem with no easy answer, whereas the existenc of any traditional notion of free will is obviously impossible once one accepts the plain reality of a deterministic universe. And so free will compatibalists can’t help themselves but to try and defend their ideas out of some irrational attachment.

  40. I can agree with the above commentators. I read virtually all of the postings, but comment on few. You’re shooting in the dark certainly, but not all of the recipients are saying ouch. Don’t stop sending them out!

    I haven’t read all of this (whew) but it does make me wonder about the GMO claims of gene incorporation by ingestion. Then again, I haven’t felt inclined to burst into seed, moo a lot, or swim in the ocean (actually, since I’m a diver, that might be an example of gene transfer).

    1. Since you are a diver, that you are obviously an atavistic example of the ‘aquatic ape hypothesis’; a hypothesis that is unfairly maligned by scientists just because there is no evidence for it! 😉

  41. The article keeps saying that HGT doesn’t invalidate evolution. But evolution wouldn’t even be invalidated if HGT were so common that we couldn’t even claim that such a thing as species existed.

    The only thing that would invalidate evolution is if HGT and any other source of genetic variation were impossible. If anything, HGT bolsters evolutionary arguments.

    1. No, I said the article doesn’t invalidate the MODERN SYNTHETIC theory of evolution, in which there is a branching tree of life and mutations are the usual source of genetic variation. But even if HGT were common, it wouldn’t invalidate evolution itself, as you correctly note.

  42. I read this and Matthew’s article with great interest, but do not feel qualified to comment. Please do not be discouraged or consider stopping this valuable service.

    1. He may be right. That argument is going to be quite technical, but I’d like to see it fleshed out rather than just dispariaging the authors qualifications and waving his hands at some general areas of inadequacy: reliability of trees, lack of quantification of gene loss, etc. It’s obvious that these are weaknesses of the paper, but without some more technical discussion it’s difficult to say whether these weaknesses merely reflect a lower quality paper than it might have been or whether it is likely that the main results reported in the paper are completely wrong.

      Hopefully someone, Eisen or someone, will take up the challenge of laying out the counter-argument in detail.

      (P.S. I think presenting this paper as a debate between HGT proponents and gene loss proponents is an example of a way that would generate more comments… To get even more comments, present the ID take on HGT. That is, if more comments is desirable, which is questionable).

      1. With this kind of paper, the devil is in the detail. Part of the problem is that the paper itself is not very clearly written and the methods are a little opaque and appear to lack the necessary controls. I’ve not yet dug into their methods in the required detail to be certain that they are flawed but they certainly appeared flawed at a first read of the paper.

        I guess Dan’s point in drawing attention to their lack experience in this area is that (given the audacity of the claims) it shifts the starting point of interpretation from “I don’t understand why they’ve done that but they must know what they’re doing so I should expend a lot of time and effort to find out” to “I don’t understand why they’ve done that and there’s a good chance that they don’t either”. Again, this does not mean that they are definitely wrong but it means that further insight is likely to come from independent analysis rather than digging deeper into what these authors did, if that makes sense?

  43. Very interesting stuff, really makes me wish I could go back to grad school for a PhD. Seems to me there are a bunch of potential projects there. For example, in Additional File 3 (line 15) they list Microtubule-Associated Protein 6 which is a class C and came from Bacteria. Bacteria don’t have microtubules so it must have been performing a completely different function originally. I think it would be fun to trace the origin of its new function.
    I also think there are probably some number of transferred genes that haven’t acquired a function and will be lost to mutation sooner or later. It would be nice to track those down.

  44. Please do not cut down on the science posts, it’s the only thing I come here for. I imagine they get fewer comments than the ones about religion and free will because people feel like they should have some actual knowledge of the subject before posting anything. If you wanna comment about religion, all you need is a strong opinion.

  45. When we hear that something ‘invalidates evolution’ its inevitable that someone will suggest that that supports ID or creationism, but that really shouldn’t be the case. The issue is not: point-mutations in isolated lineages vs. HGT, endosymbiosis and other esoteric mechanisms. The issue is natural mechanisms vs. God

    1. “The issue is natural mechanisms vs. God”
      Since there are a huge number of gods and not all loony fables that “explain” life on Earth involve alleged gods, I would generally prefer to say something like:

      “The issue is natural mechanisms vs. evidenceless mechanisms like magic or alleged outer space aliens.”

      The former covers all gods and even Scientology, while the latter covers the “naturalistic” explanations that require visitations from outer space thingies that always seem to be bothering some small part of humanity.

  46. “I hope so, because I’m worried, in view of the dearth of comments on science posts, that people are skipping them. ”

    I read them. Especially the ones like this one, which explain recent developments I’d never run across or understand otherwise. And I track back. I have read every post going back years on group selection for instance. The best antidote to woo is clear thinking on science, which your science posts provide.

  47. I too try to read the more sciencey posts, but often have nothing to say.

    Here, I wonder – what’s the *mechanism* for HGT in these cases? I understand how viruses, which coopt the biochemical “machinery” of their hosts might do it. But wouldn’t most cells from (say) a fungus in a human (or a fly) be just phagocytosed by the immune system or ignored?

  48. So if cat breeding for desired traits is artificial selection as opposed to natural selection, is there a correlation between HGT (which is natural) and gene splicing (for GMOs for example) which is artificial HGT. I guess I’m asking if the knowledge of HGT spurned research into gene splicing or gave a framework or model. It’s probably apples and oranges…

    Relating to the first paragraph, I always read the science posts but don’t always comment. Sometimes I feel a question looms (like this time) and sometimes I’m just happy to learn something new and read the comments to see what others add or come up with.

    Echoing many readers here- your science posts are my favorite posts. These are the posts from which I truly learn new findings that I wouldn’t be able to locate in the proverbial needle of scientific haystacks that are out there. Plus you and Dr. Cobbs’ ability to extrapolate and simplify (not dumb-down) the material for the non-scientist is distinctive and extremely important to me.

  49. I generally read the scientific articles but don’t have anything to comment.

    Except in this case – I gave up my subscription to New Scientist because of the “Where Darwin Went Wrong” cover and article. Still not a comment about the science though.

  50. Thank you. And keep the science coming!

    I have been wondering about HGT and its place in evolutionary theory for a while now and this addresses a lot of my questions. Some articles I have read were rather sinister in their implications. This seems like a more reasonable perspective.

  51. Ok so I will try to comment without sounding too laypersonish( do you see we are all just a bit intimidated by your research and that’s why we refrain from commenting?)

    If this genetic material is absorbed mitochondria-like into the genome what seems to be the mechanism of uptake? It seems a long path to travel from the ingestion of protist to the gamete of a primate in order for the genetic material to be passed on to the next generation.

    1. This is an excellent question, of course. I vote for a follow up post about possible mechanisms!

      Of course, one kind of mechanism is for the destruction of the cell walls of an infecting agent, releasing it’s DNA into the cytoplasm of a cell it’s infecting, and the foreign DNA then makes it’s way into the nucleus and is take up by accident, perhaps during the repair of a broken DNA strand (hybridizes to a sticky end). This just-so story is just to illustrate the kind of thing that might be possible. I have no idea if it happens that way. Some bacteria transfer genetic material to other bacteria via a “pilus”, so one can imagine that such a mechanism might result in DNA being injected into nucleus of a host eukaryote somehow. Once in the nucleus, a number of different kinds of accidents might result in it’s being incorporated into the genome. And as for the gremline, that is a matter of chance, to infect an embryonic stem cell or a gremline cell. So that would mean that such incorporations into cells must be a lot more common even than we see since only a small fraction will ever happen to be in the gremline, and a smaller fraction still will get fixed in a population.

      So many interesting questions! This should be the beginning of an 8 part series!!

    2. I don’t think Jerry meant that eating something was a potential mechanism in animals. That seems unlikely by any standard and to actually be moved to the germ cells would be even rarer.

      There are three main ways I’m aware of that result in HGT.

      Bacterial transformation. Some bacteria are capable of picking up naked DNA from the environment.

      Bacterial conjugation. It’s basically sex for bacteria but certain bacteria and also create the conjugation structures and send DNA into eukaryotic cells.

      Viral transduction. This is most likely if you’re talking about humans and multicellular organisms. Some viruses are able to integrate themselves into their hosts genome. That’s why you could have HIV for years with no signs. It also makes it impossible to cure some viral diseases. Even if it’s not active, the virus exists inside your DNA. It’s possible for mutations to trap viruses so they can no longer get free.
      In addition, when viruses leave cells to infect new cells they package their own DNA into little capsules. It’s possible that a capsule will not get viral DNA but instead gets the DNA of the current host and then delivers that DNA to the next cell that it infects.

      You can see illustrations of these methods in the slides of a talk I gave.

  52. Have not had time to read the comments above yet, but fascinating that the ABO group (its really just the A and B group, since type O is a non-functioning mutation) came to us by HGT. I know that these are shared between us and chimpanzees. I wonder where the original came from.

    1. It’s hard to tell from the paper, because the taxonomic coverage isn’t good. Such information as they present comes in figure S1B (in the supplementary information). There, they show nearly identical proteins in a bunch of mammals. They don’t say if they looked at any other vertebrates, but it seems to be absent in ecdysozoans. And suggested homologs, though very different in sequence, occur in some protists and bacteria. That might be evidence that the transfer happened from protist or bacterium to some ancestral mammal, or it might just show that they haven’t sampled enough species. I wish there were some indication of all the places they looked and found nothing.

  53. Professor – just one data point re your comment on readership of science posts: I read them all, but comment less on them than, say, posts on theism. Not at all because science posts are less interesting, it’s just that the theism/politically-oriented posts trigger more discussion and emotional (if not comment-worthy) response.

    So please take heart! I for one place extremely high value on these posts – there are few or none of this depth and quality in the blog-o-sphere. And I’m sure I have plenty of company in the readership here.

  54. Just wanted to add a comment saying thank you for taking the time and effort to write up this explanation. I read WEIT for the science – everything else is a bonus!

  55. 0.2%- 1.5% means that it is quite rare, as one would expect.

    Thanks this is a fascinating paper that I certainly would not have delved into on my own.

  56. I remember being amazed when reading Dawkins’s The Ancestors Tale and the description of the way sea slugs eat hydrozoa (small jellyfish) and then transport the hydrozoan nematocysts (containing the sting) to their own body surface for protection. Could this result in HGT?

    1. A strong claim has been made about a similar system in which sea slugs eat and then transport and store chloroplasts from algal tissue.


      That claim has been strongly criticized, and is probably wrong, but interesting to think about.


  57. A fascinating read. Not up to adding anything by way of comments other than thanks and keep them coming because they are appreciated.

  58. I am most interested in the scientific posts. Keep writing.
    I do not know enough to comment. Hereafter I will say that I read them.
    I am so old that one of my profs in grad school (anthropology) insisted that viruses did not exist. He thot viruses were an excuse that doctors made up when they did not know what caused a disease.

  59. I am (or was) a trained scientist (developmental genetics mutating toward evo-devo in late career)but still find I seldom have anything to add to your well-written science posts. Incidentally, in addition to interesting new work, I like to see critiques of bad science.

    1. Your career seems parallel to mine, btw. My route was entomology –> developmental bio –> evo-devo. Hard to make a living at it without being a big grant writer, but a wonderful experience anyway.

  60. Jerry,

    I may be the opposite of some people here, but I SKIP THE CATS and read the science based articles. I really appreciate the evolution and other more science based articles. I may not comment but it is only because I use them as a resource to learn and not a source to comment in agreement or disagreement.

    Please keep them coming…

  61. Late to the party; haven’t read the comments.

    I know this borders on a Da Roolz violation…but I suspect I would likely stop frequenting this Web site were there to be a precipitous drop in the science posts.

    For example…my mind was just blown at the thought that primates have plant DNA as a recent addition, something I would have simply assumed was impossible. Now I need to go find what those genes are, if it hasn’t already been mentioned in the comments above that I haven’t yet read….


      1. Yes — but that’s only after the DNA gets in there in the first place. Once it’s there, of course, the cells treat it like any other stretch of DNA.

        What I want to know…is how did the plant DNA get inside a primate cell in the first place?


        1. Viruses seem the most likely vector to me. They can pick up genes from their hosts and carry them to new hosts.

          Aside from that, while it’s not too difficult to imagine DNA making it into the bloodstream, I’m not seeing how it would make it into germ line cells. I think it would ultimately still have to be a viral vector, just not necessarily for the whole trip.

          WAG, of course.

  62. I’m with Ben on the plant genes!

    Also, love the science posts. Likely too dim to grok most of them, but the very act of reading them keeps the fires going.

  63. Does anyone with more expertise than I in this subject have a handle on the mechanism of HGT in higher eukaryotes? The most plausible mechanism I see in my perusal of the literature is by viral vector, particularly retroviruses. [Note that protists, archaea and bacteria, probably because of their relative simplicity are quite promiscuous gene sharers by other mechanisms.] Gene transfer from endosymbionts (like Wolbachia, a parasitic nematode) have been demonstrated. There is apparently some debate about how often (if ever) HGT occurs in chordates.

    1. Those are the most likely mechanisms. One thing to remember about this paper is that they are only looking at HGT between kingdoms. This is much harder to explain (except from intracellular pathogens/symbionts), which is probably part of the backlash against the weak methodology. Even if most of their HGT predictions turn out to be false, there is probably a lot more HGT going on between more closely-related organisms that share parasites and pathogens etc. The chances of genes getting into the germline is incredibly small. However, the number of reproducing animals is incredibly large. The ultimate probability of HGT from this incredibly large number of incredibly unlikely events is anyone’s guess at this point.

  64. I hope the number of comments completely demolishes any idea that science posts are unread!

    I too, could not possibly comment on most, (all really) but I do want my vote counted. They are the best – after cats of course

  65. I only really comment on science posts when the study in question has marked flaws that I’m able to spot, or when there’s a question just pleading to be asked.

    Over the last couple years, however, Jerry (and Matt and Greg) have almost exclusively either covered well-conducted studies (the flaws of which are either absent or beyond my ken), or have preemptively pointed out everything I might have addressed in a comment.

    So do a shoddy job and you’ll get more comments. Or, keep doing yeoman’s work on the science posts, and choose to see the relative dearth of comments as evidence of a job well done.

  66. Great post. I read every one BTW, from Hili to science to anti-theism posts. I like the balance of variety. Very glad to have discovered this amazing community of people (and felines). Regards….

  67. I don’t understand how genes from one organism get into the germline cells of organisms that reproduce sexually. (In other words, how does the DNA from a plant get into your gonads.)

  68. I wanted to let you know that your science posts are very much appreciated, at least by me. The papers you review can be very technical and you are able to distill the essence in a way that is easily grasped by your readers. I graduated college with a degree in molecular cell biology but never had the opportunity to work in my chosen field. Your blog is one way I work at staying informed of advancements in the sciences. Thank you.

  69. To concur with all the comments above. As a lowly, hard-working evo biology PhD student I get a tremendous amount out of the science posts but don’t always have the time to post any immediately thoughtful responses, but rest assured these posts stimulate the old grey matter and your efforts are much appreciated!

    PS- read the Okapi post with especial interest as I realised one of my friends was an author on that paper 🙂

  70. I was a trained scientist and I read all the science posts here. I don’t usually comment on them, however, for several reasons. First,when I was a working scientist, my field was muscle structure, function, and disease, and I taught human gross anatomy. Second, after a couple of bad personal years, and a resulting loss of grant funding, I became a research grant administrator for several health agencies. Third, I’ve been retired for 10 years.

    I try to keep up on science by reading Science, Nature and this bl*g. I’m fascinated by genetics and evolution, but since my research was in a very different area, I seldom feel I have anything to contribute to the science posts here. But I appreciate them very much and I’m pleased to learn from this one that I might have some archaea genes somewhere.

  71. So Spiderman could be real?

    I liked this science post but I don’t agree on “173” to be “hundreds”. “hundreds” is 200 at minimum.

  72. Terrific Jerry, just like being in one of your classes. I’d comment more but being out on the West (Left) Coast it’s late at night when I get around to accessing this site so I tend to rush through the posts. When I do comment, I’m always surprised to find (the next night) that some people comment even after I do. In any case, don’t stop the science posts.

  73. I want to preface this comment by saying that HGT is very real and very interesting… BUT… as a bioinformatician who works with genomes, gene families, alignments, trees and homology searches on a regular basis, I am not convinced by the robustness of the methodology of this paper.

    The authors make some rather strange decisions (using bitscore differences, for example, which are highly biased by gene length and repeats) and do not seem to sufficiently consider incomplete sampling and gene families. For their phylogenetics, for example, they only consider a few top BLAST hits. Lineage-specific duplications and losses could mess up perceived relationships. As pointed out above, without seeing the alignments, one cannot tell how good their trees are, and without seeing what BLAST hits they are excluding one cannot judge how easily they may be mislead. For only 173 proteins, they should have full phylogenetic analysis and alignments available in supplementary data. Ultimately, only detailed phylogenetic analysis can robustly identify HGT.

    My final major gripe with the paper is that I do not think that they sufficiently consider contamination and genome quality. They consider contamination of the target metazoan genome but not, as far as I can tell, the non-metazoan genomes from which they propose the genes to have come. This is despite explicitly screening some of these out for being *too* contaminated. Likewise, although they use “complete proteomes”, there is no such thing. Sequencing, assembly and annotation issues can all result in proteins being erroneously missing, particularly if a gene has ended up near a telomere or other difficult-to-sequence region. Individually, the probability of error might be low but when you start looking at 20,000 genes and add genuine gene losses into the mix, the likelihoods become higher.

    This is not to say that the results are definitely wrong… just that they have fallen rather short of being convincing. Maybe chance results are still unlikely despite the issues. Maybe not. However, the more fantastic the claim, the more robust the supporting evidence needs to be! I am sure that some of the reported HGT is real – particularly those previously reported – but I would definitely need to see more convincing evidence for their new ones.

    My final comment is just to point out that these are only the protein-coding genes – humans have a LOT more than 20,000 genes!

  74. Viruses and intracellular bacteria (such as Woolbachia) are probably the most likely source of HGT in metazoans as they can get into the germline. However, it should also be noted that they can also very easily contaminate genome sequences and assemblies. Looking for assembled neighbouring genes from the host (as the authors did) is not enough; the assembly algorithms themselves can make mistakes. You really need to perform PCR or targeted resequencing on germline tissue to demonstrate that your bacterial/viral gene is integrated into the host. (Although the original human genome was sequenced from sperm, I do not think that is true for most metazoan genomes – I suspect that we are actually looking at potentially contaminated/infected somatic genomes in most cases.)

  75. My guess is that most of us came here originally for the science posts, though we each keep coming back — my guess, again — for our own reasons (making us a modern, blended family per the Tolstoy dichotomy).

    I almost always read the science posts, but refrain from polluting their comment sections with cheap wisecracks or gormless inquiries flaunting my ignorance.

    Also, I never pass on the opportunity to be(according to the running comment count above) the 186th “me, too.”

  76. Very interesting, thanks for summarizing! I do often read the science posts. I’ll try to make an effort to contribute more 😉

    “by eating of one organism by another, and then incorporation of that organism’s DNA into the genome” – I see others have made similar comments, but I can imagine anti-GMO people latching on to this kind of statement. Presumably this is a non-issue and would reflect a misunderstanding of some kind, but at the same time, I feel ill equiped to explain why consuming GMO food isn’t a problem in light of that sentence quoted above.

    1. There are two ways to look at it:

      1. GMOphobe: “See! These genes can just hop from one species to another. Anything we engineer into our crops could escape and spread into weeds, pests etc. or even us! Frankenfoods! Aaaargh!

      2. GMOphile: “See! Taking a gene from one organism and putting it another is actually quite natural. We are just accelerating natural processes with GM.”

      Both are true to some extent. GM is arguably no less “natural” (and *much* more efficient) than extreme selective breeding. In *theory* most modifications could happen “naturally” albeit only with infinitesimally small probabilities. Likewise, it is theoretically possible for genes introduced by GM to “escape” and cross into other species via HGT or hybridisation. The important questions are: (a) how likely is a particularly HGT event? and (b) what difference would it make if it happened? The answer in most cases for GM foods is (a) very unlikely and (b) nothing of consequence, but obviously this needs to be considered case-by-case.

      The chances of GM bacteria giving their new genes away to all and sundry is much higher, but we generally modify bacteria to do something useful (bioremediation etc.) so it would not necessarily be bad for it to “get out”. (And whether there is the required selective environment to maintain it is another question again.)

  77. As a scientist, I can say that I still get a lot from many of your science posts. This one is particularly interesting – and it seems reasonable that metabolic processes should ber involved in most HGT cases. The fundamental molecular machinery of metabolism is what we would expect to be most universal. It seems reasonable that the odds that a transferred gene from a very different species would provide a favorable adaptation would be greatest for such cases.

    As a chemist, I don’t much peruse the bilogy literature. I would miss this if you didn’t post on it.

    1. Unfortunately, the metabolic enrichment is another example where the authors provide insufficient detail of their methodology and, without sufficient controls, might be misreporting bias as enrichment.

      By definition, all of their HGT genes have detectable homology in at least two kingdoms. As you correctly point out, fundamental machinery is more likely to be conserved over such distances. Therefore, they could just be picking up this bias, rather than a specific bias of HGT genes.

      They should compare their HGT genes for enrichment versus all genes with inter-kingdom blastx hits above their threshold. As they do not state this is what they did, one has to assume that a “default” analysis against all genes was used. It is therefore impossible to say whether their enriched gene ontology terms have anything to do with HGT or simply reflect the types of genes that are most conserved (and, to a lesser extent, better annotated) between kingdoms. If that makes sense?

      (The more I read this paper, the more cross I get. The reviewers and/or editor did a terrible job at making sure that it had sufficient detail and rigour for publication, particularly in a decent journal like Genome Biology. I hope the results don’t turn out to be as wrong as I fear.)

      1. By definition, all of their HGT genes have detectable homology in at least two kingdoms. As you correctly point out, fundamental machinery is more likely to be conserved over such distances. Therefore, they could just be picking up this bias, rather than a specific bias of HGT genes.

        As in, these could be ancient genes that haven’t been subject to modification from evolutionary pressures since common ancestry? Or perhaps even a bit of convergent evolution that solved the same problem with the same sequence sometime after divergence?

        I did wonder about that….


        1. Convergence is another issue that they don’t adequately consider. I’ve not finished digging into the data but it could be part of the explanation.

          I was more referring to observation bias. Imagine you are looking for people wearing hats in the cinema, from the back. You spot a bunch of heads, some of which have hats. You then grab only the people with hats, measure them, compare them to the general population and conclude that people who wear hats are above average height…

  78. I read the post about Okapi, but was busy at work yesterday and haven’t yet read this post. Keep the science coming and I will eventually read it.

    My focus at university was plasma physics and remote sensing; so I don’t feel that I have much to add to biology posts. But that doesn’t mean I don’t read them and find them fascinating.

  79. I love the science posts. In fact, it’s the only reason I follow the blog. I very quickly skim through the non-science posts (even though I agree with most of the religion posts; however, there’s not much for me to learn there anymore). I’m a scientist, and there aren’t really that many blogs with high level science content, so I definitely appreciate the work that goes into it. I never comment; that’s just something I’m not that interested in doing. Keep up the fantastic work!

  80. I enjoy reading it…. And it is written by an expert so I pretty much can accept it without having too much doubts. :)….

  81. I truly enjoy and do read the science articles. I learn a lot from them, but just don’t have the expertise to offer useful comments. FWIW, I do have a neighbor who knew Lynn Margulis. We nerds do talk to each other about what we’re reading.

  82. Dear PCC
    Yes I read it. I read 95% of them and like now I am reading it in my back yard where I am drinking gin and typing on a ridiculously small keyboard. AND I don’t have my reading glasses (not that they would help). Please forgive me. That’s my excuse and I’m going to ‘sticky wicket’.
    Thanks for keeping me smart 😉

Leave a Comment

Your email address will not be published. Required fields are marked *