Barcoding redux 17 Mar 200818 Sep 2017 So, here I am in Phoenix airport, waiting to go back home, and I read T Ryan Gregory’s snark about me and barcoding. Apparently I am to learn only from his blog posts and not from (perish the thought) critics. One should never attend to critics. My crime was, of course, to say that I thought Brent Mishler of UC Berkeley and others (including mein host in Phoenix, Quentin Wheeler, and Kip Will) were correct in their concerns that barcoding was being touted as a replacement for proper taxonomy and that it will draw resources from it. What are the issues? There are three, as I see it. One is that a novel technique ought to be proven before being widely adopted. This is pretty obvious, as work done using a flawed technique will waste time, introduce errors, and cause, as it is this time, all kinds of territorial disputes in the sciences. The territories here are between “organism” biologists, some of whom go by the term “comparative morphologists” and others who go by the label of “field biologists” & c., and molecular biologists. Molecular biology has been enormously successful over the past thirty or more years, but it does tend to hegemony. Many biology departments lost all but a rump of other disciplines over these years, as Ed Wilson complains in his biography, Naturalist. The second is that a technique ought not to be used for more than it can achieve. Suppose barcoding does something very well – it should not therefore displace other techniques that do other things. But this is historically what happens. To a man (it’s usually a man) with a hammer, everything is a nail. The third reason is sociological. Scientific disciplines are fighting over the same territories and resources. If you can get funding with a new technique, you can expand departments based on that technique, and other departments or traditions will lose out. It doesn’t matter if this funding is “new” or existing – there is a limited amount of funding for science and it often turns out to be a win-lose. Worse, funding can cause corruption in the assessment of scientific techniques – industries that make the tools will, as we have often seen in medicine, promote those technologies to the detriment of actual scientific merit. Now what the critics to whom we should not listen are saying is that there are concerns in all three ways. The first is that barcoding has not been tested. A basic assumption is that the COI gene correlates with species, or some kind of biodiversity, or something. But so far, this hasn’t been empirically tested for more than a few, rather atypical, groups. In fact I am told (but have no personal knowledge) that the evidence is to the contrary – COI is a particularly bad indicator of species. If this is true, why is barcoding being so strongly promoted? It has the advantages of simplicity, speed and “economy” (but more about that in a minute). Of course if the product is epistemically suspect then speed, simplicity and low cost are irrelevant for the science (but not for the careers and profit margins of those who promote it). I wonder if this is popular largely because it can be sold to politicians and funding agencies. The second criticism is the one Mishler mentioned in his talk: it displaces taxonomy. Gregory says it doesn’t, and that this is money that would not have been available otherwise, but those who are trying to get real taxonomy off the ground again, before biodiversity is all extinct think otherwise. And even some advocates of the barcoding project say that this replaces actual taxonomy – at the Second International Barcode of Life conference in Taipei in November 07, Les Christidis of the Australian Museum stated that replacing taxonomists with barcoding is an economic measure (see his Powerpoint slide near the end). As Christidis is one of the biggest employers of taxonomists in Australia, this gives the lie for one country at least. I haven’t had time or bandwidth to check other presenters, but it sure looks to me like at least some barcoders see this as a taxonomy replacement. The list of barcoding partners looks like some of the museums that traditionally have employed taxonomists – are they going to somehow be different from the Australian? And this also supports the third point; this is a cross-disciplinary pissing match. Molecules are seen with such awe by some that they are the magic marker for all that is important. In the philosophy of biology, this has been called “gene-centrism” and the subject of considerable criticism for many years, but it is wider than genes – proteomics also seems to be seen as some kind of “final answer” to vexing problems. Now molecular biology is a really useful part of biology, but it isn’t all of biology by any means. It is what it is, and no more. If tracing a particular gene is shown to be important (and we’ve been through this before with the use of another mtDNA sequence, 16s rRNA, where it turned out that initial expectations of utility were overdrawn) then do it, but don’t say or imply that it does more than it is shown to be good for. I won’t return to this because philosophically it isn’t interesting, despite the sociology. It is a return, in philosophical terms, to a mode of artificial classification by single keys that was briefly thought to give the structure of the natural world. We know now, or should do, that is not true – we need lots of information to do this in biology (and this post by Ebach and Williams discusses that problem) but only that which actually informs. Ecology and Biodiversity Evolution Social evolution Species and systematics
Evolution Ontologies of biology 6 Nov 2008 I am keen to jot down whatever I can about the ontologies of biology – not just evolution, but also molecular, developmental, taxonomic, ecological and other domains of biology. I want to do this in a relatively systematic manner, so I would appreciate readers noting in the comments the sorts… Read More
Evolution Sociobiology 5: What is at issue 22 Nov 200718 Sep 2017 So now, I think it’s worth asking what we really can achieve by doing sociobiological investigations, and some of the traps in previous attempts. Read More
History What is a species? 29 May 2008 If somebody asked me to write a short essay giving an overview of my favourite topic, the nature of species, I doubt that I could. I can write a long essay on it (in fact, several) but it would be excruciatingly hard to write a short one. For that, we… Read More
Apparently I am to learn only from his blog posts and not from (perish the thought) critics. Actually, I merely commented that you receive most of your information about the topic from critics rather than from those who do the work. The territories here are between “organism” biologists, some of whom go by the term “comparative morphologists” and others who go by the label of “field biologists” & c., and molecular biologists. This is incorrect. Most of the people I know who are interested in barcoding have a strong interest in organisms and diversity and do field work themselves and/or collaborate with field biologists and taxonomic specialists. Suppose barcoding does something very well – it should not therefore displace other techniques that do other things. But this is historically what happens. Indeed — barcoding complements morphology, behaviour, ecology, and other types of information; criticizing it because of what has transpired in other cases hardly seems logical. I wonder if this is popular largely because it can be sold to politicians and funding agencies. You have no idea the challenges that were involved in getting it off the ground. The first is that barcoding has not been tested. A basic assumption is that the COI gene correlates with species, or some kind of biodiversity, or something. But so far, this hasn’t been empirically tested for more than a few, rather atypical, groups. In fact I am told (but have no personal knowledge) that the evidence is to the contrary – COI is a particularly bad indicator of species. If you don’t know and have not examined the data then why do you make the claim? Have you read any of the literature (data papers, not criticisms) from the past several years? As of this moment, the database has 359,366 records from 37,177 species, including plants, fungi, protists, vertebrates, and invertebrates. The second criticism is the one Mishler mentioned in his talk: it displaces taxonomy. Gregory says it doesn’t, and that this is money that would not have been available otherwise, but those who are trying to get real taxonomy off the ground again, before biodiversity is all extinct think otherwise. And even some advocates of the barcoding project say that this replaces actual taxonomy – at the Second International Barcode of Life conference in Taipei in November 07, Les Christidis of the Australian Museum stated that replacing taxonomists with barcoding is an economic measure (see his Powerpoint slide near the end). As Christidis is one of the biggest employers of taxonomists in Australia, this gives the lie for one country at least. I haven’t had time or bandwidth to check other presenters, but it sure looks to me like at least some barcoders see this as a taxonomy replacement. I encourage people to see the slides from Christidis’s presentation, which are available online (“Number of specialist taxonomists may be less but number of generalists taxonomists will increase many-fold”). And this also supports the third point; this is a cross-disciplinary pissing match. Molecules are seen with such awe by some that they are the magic marker for all that is important. No it isn’t. Barcoding networks include individuals with molecular expertise, field biologists, taxonomists, and others. Hardly any barcoders are in it just for the molecular data, most are interested in systematic, evolutionary, ecological, and other topics. The goal is to provide a way to obtain species identifications that allow individuals who are not taxonomists to access the important knowledge that taxonomists produce, and to assist with the description of new taxa in a more efficient way. I, for one, think taxonomy is very important.
Don’t go lumping us molecular biologist in with the dirty molecular ecologists. They may work with molecular markers, but they’re still crunchy ecologist who don’t shower.
Don’t go lumping us molecular biologist in with the dirty molecular ecologists. They may work with molecular markers, but they’re still crunchy ecologist who don’t shower.
Don’t go lumping us molecular biologist in with the dirty molecular ecologists. They may work with molecular markers, but they’re still crunchy ecologist who don’t shower.
Hey some of us shower once every single week! But in terms of barcoding, it is my opinion that the anti-barcoders are continuing to recycle arguments that I’ve been hearing for the last 4 years, while those who actually use the technique are continuing to recognize and address weaknesses to improve implementation of the approach and interpretation of results. I don’t do “barcoding” (I really hate the word actually) but I know several people involved in such projects, and not a single one of them thinks of the resulting COI data as an end goal (and in most ongoing projects I am aware of, more genes than COI are being used to verify results). Noone is trying to replace “real taxonomy” (whatever that is). If anything barcoding is likely to highlight situation that are most in need of examination by taxonomic experts (hyper-diverse but morphologically cryptic groups) that might otherwise be overlooked.
To a man (it’s usually a man) with a hammer, everything is a nail. For a woman, wouldn’t that be a file? *runs and hides*
To a man (it’s usually a man) with a hammer, everything is a nail. For a woman, wouldn’t that be a file? *runs and hides*
To a man (it’s usually a man) with a hammer, everything is a nail. For a woman, wouldn’t that be a file? *runs and hides*
To a man (it’s usually a man) with a hammer, everything is a nail. For a woman, wouldn’t that be a file? *runs and hides*
To a man (it’s usually a man) with a hammer, everything is a nail. For a woman, wouldn’t that be a file? *runs and hides*
Just as an aside (this is material from undergraduate theses done by students in our lab in Singapore), while COI isn’t particularly bad (no worse than cytochrome B or ND1 in birds), it seems to have a steady success rate of about 75%, no matter whether one looks at all metazoan sequences in Genbank or in my case, a comprehensive sample of widespread species in a single family of Diptera (Sepsidae). The errors were mainly due to 1) singletons sequences in the database, which can be improved with better COI sampling, or 2) shared barcodes between very different species (I had a Platytoxopoda sp. identified as a Scathophaga stercoraria with 100% confidence on BOLD-IDS), which cannot be improved no matter how much of life is barcoded. Also, a predetermined COI distance (3%) may sort out most of the sister species out there, but will fail in the case of very old and very young species.
Just as an aside (this is material from undergraduate theses done by students in our lab in Singapore), while COI isn’t particularly bad (no worse than cytochrome B or ND1 in birds), it seems to have a steady success rate of about 75%, no matter whether one looks at all metazoan sequences in Genbank or in my case, a comprehensive sample of widespread species in a single family of Diptera (Sepsidae). The errors were mainly due to 1) singletons sequences in the database, which can be improved with better COI sampling, or 2) shared barcodes between very different species (I had a Platytoxopoda sp. identified as a Scathophaga stercoraria with 100% confidence on BOLD-IDS), which cannot be improved no matter how much of life is barcoded. Also, a predetermined COI distance (3%) may sort out most of the sister species out there, but will fail in the case of very old and very young species.
Like Ryan, I have been involved in writing a grant to fund a DNA barcoding project. Without giving away too many details (we are still trying to get funding), I will say that in the current version of our grant proposal the amount budgeted for the work of “real” taxonomy and systematics is nearly four times that budgeted for the molecular biology. By “real” taxonomy and systematics I mean the usual work of taxonomists and systematists (both researchers and curators), that is, projects involving collection and (morphological and molecular) analysis of specimens in order to ask important questions in systematics.
The sociology surrounding molecular triumphalism (of which barcoding is one manifestation) is actually not unique to biology. I’ve just finished reading “The Trouble with Physics” by Lee Smolin. According to him, proponents of string theory have completely taken over the hierarchy (and funding) of modern physics which has led to 30 years of stagnation. Reading his book was comforting and disturbing. Comforting to know that taxonomists aren’t the only ones banished to the unfunded wilderness, but disturbing to realize that the distortions in biology are propagating everywhere. Is barcoding the string theory of biology?