Why the current DNA barcode idea is a bad one
Jun 30th, 2009
Okay, I have been debating whether I should post this or not but here goes. I have been involved with the DNA barcode idea through a series of strange events and communications, most of them from rather irrational proponents of the DNA barcode idea – including calling me a ‘reptile brain’. Hmm. Okay, sounds like a reasonable argument.
But I saw Brent Mishler give a talk in Singapore at the Willi Hennig Society meeting and I was very glad to see someone prominent with a similar view on the topic and present some very nice and clear arguments.
Like Brent, I would like to clarify one thing from the outset. I am not against using DNA to identify things or to classify taxa. This is an old idea and has been used effectively in a wide range of settings. Brent quoted someone (I forget who, sorry) in relation to DNA barcoding: “Your idea has some new ideas and some good ideas. Unfortunately, your new ideas are not good and your good ideas are not new.” Precisely. I am pushing the use of DNA fingerprints for a wide range of applications but I envision the ‘barcode’ to consist of a set of thousands of loci. The selection of these loci will depend upon the relatedness between potential species. It will not rest on the arbitrary selection of 2-3 loci.
The reasons I do not like the current form of the DNA barcode idea can be broken down into two sections: one relates to Public Relations and the other to Biology.
Public Relations
1. Species are not a can of tomato soup. This is directly from Brent’s talk. A DNA barcode implies that all tigers are the same. All humans are the same. You buy a can of soup and the ingredients are on the outside and every can of soup is exactly the same. This implication goes against everything we know and is fundamentally a creationist idea. It is a very bad message to the general public and to politicians.
2. By promoting the idea that we can identify species, using some magic figure like 2% difference in DNA sequence based upon a tiny tiny fraction of the genome, makes the science of biodiversity and speciation seem very trivial. While absolutely no consensus has ever been reached on the question, “What is a species?” – the DNA barcode group claims they can differentiate everything based upon this trivial and meaningless definition. The idea is an insult to taxonomy and biology and will undercut the real research into the basic question of ‘what is a species’.
3. From my interactions with the DNA barcode folks, they act like a cult. They do not respond to sincere and critical discussion. They flame you for criticising them. They coerce others into agreeing with them. I was disappointed when I read a paper by a good friend of mine, someone who I know knows better, where they tested a range of loci to identify a major group of organisms. Actually, one locus other than COI performed much better but COI did okay, so the discussion ended on a compromised note – we have a better locus but the ‘barcode’ is okay. Why choose a locus when a better one is available??? unclear — but if you think of them as a herd, a bandwagon running over everything in their path, yes, it makes sense then.
Biology
1. There is no magic locus that can identify ALL species. This is just a simple fact. As a botanist, I have seen direct evidence and know for a number of colleagues that the same locus evolves at a very different rate in different lineages. Makes sense to me, unless you proscribe to a pre-molecular evidence concept of a reliable and consistent molecular clock.
2. There is no bioinformatic argument for using a single locus or even 4-5 loci. This is a pre-internet, pre-desktop computing idea. There is no reason why each group, studying different organisms, should not develop and define their own best and most informative locus. At the meeting in Singapore, some people were astounded that the COI region in flies is often far above the mythical 2% level of sequence divergence, even within a species. Well, two things: A) they are flies – very short generation times and rapidly evolving – I would expect them to diverge much more quickly than say a rainforest tree and B) different loci evolve at different rates in different organisms, for no good reason, it is random.
3. The barcode is a one trick pony. It provides a single piece of information and when it is wrong, it provides no way of solving the problem. So, when you really want it to work, in diverse genera of closely related organisms, it’ll provide no answer.
4. The single locus PCR analysis will soon, if it is not already, be obsolete. Not only is there no reason to look at a single locus from a bioinformatics perspective, there is no reason to look at a single locus from the DNA sequencing technology perspective. We can already easily sequence the entire mitochondrial genome for all species. It would not cost any more than the budgets I have seen for DNA barcoding projects. They are investing a tremendous amount of money in a technology that will soon be obsolete.
5. Abundant evidence exists already that across a broad spatial scale, the same species can possess different cytoplasmic genomes. There are clear theoretical models explaining how this happens through introgression and how this has a relatively trivial effect on the overall level of species divergence. The DNA barcode concept is anti-evolutionary and simple goes against theory.
Okay, there, got that off my chest.
