Wednesday, February 28, 2007

DNA barcoding and "15 new bird species:" Not so fast

The popular media and internet have been buzzing recently about the supposed discovery of 15 new species of North American birds based on the DNA barcoding results by Kerr et al (go to Google News for "birds" and "barcoding" for a plethora of articles). To borrow a phrase from our British friends, bollocks. I'm discussing the press reporting about the study here, not the original study itself which takes a much more conservative tone. Simply put, the implication that DNA barcoding is capable of discovering new species all by itself is nonsense. There are two fundamental reasons why this is true.

First, there is much more to speciation than genetic differentiation. The concept of a species is not simply two populations that have not interbred; it is two species that are reproductively isolated by biological means. Genetic distance and species-level distinctions are correlated, but not in an absolute, deterministic sense. Isolating mechanisms and assortative mate choice are critical, and can't be determined just from genetics. You have to actually go out and study real, whole birds, not just their DNA. And in the case of hypothesized species that do not at present overlap in their ranges, you have to look at their life histories and reproductive biology and make your best judgement as to whether they would remain reproductively isolated if they did come into geographic contact. Genetic closeness also does not strictly imply that two entites are the same species, either. Reproductive isolation can be accomplished with only very slight genetic distinctions, and if the two species have only very recently become isolated there might not yet be much total distance between their genomes. Genetic distincitiveness of various populations is only one of many factors considered in splitting and lumping species; indeed, in many ways, it is primarily a marker to tell you that there might be something going on here that needs to be examined in the larger context.

On beyond this, even just within the genetic context, the particular gene being used for the DNA barcoding project is uniquely ill-suited to "discover" closely related, potentially "young" species. Cytochrome C oxidase I is part of the mitochondrial genome. While this gives it many useful characteristics. including a rapid rate of change over time, the salient feature of mitochondrial genes is that they are inherited in an ASEXUAL, purely matrilineal fashion. They are all but useless for investigating rates of gene flow between putative species, which occurs by sexual reproduction. Mitochondrial DNA (mtDNA) does not mix and blend the way the chromosomal genome does. A hybrid does not have hybrid mtDNA. It has 100% of the mtDNA from its mother, and 0% of the mtDNA from its father. In a hybrid zone you will not find intermediate mtDNA genotypes; each individual bird will have the mtDNA from either one species or the other, with no intermediates. As the fundamental, rock bottom genetic consideration when splitting or lumping species is gene flow between the two taxa, mtDNA is a poor tool for assessing this. If you already have the species defined and their mtDNA characterized, then mtDNA barcoding will let you assign any single individual to a species, or at least the species of its maternal line. That is the purpose of DNA barcoding. Discovering "15 new species of North American birds" is not its purpose, nor is it an appropriate use for it. And, I should add, neither is lumping 8 species of "white headed gulls," no matter how much the lariphobes out there would wish it. MtDNA can only suggest these things as possibilities to be investigated further.

There also seems to be a thread in the coverage of this research suggesting that the discovery of "cryptic" species is somehow a news flash. In reality, the uncovering of superficially similar bird species that had been previously lumped has been happening for decades, and genetic taxonomy has been an important part of this process. Back in the 1970's when that bloodbath of an AOU checklist supplement came out, obliterating our juncos, flickers, and many other species, there was also one change that was actually the herald of the future: the split of Traill's Flycatcher into Willow and Alder Flycatchers. Here were two species, differing in vocalizations, breeding biology, and genetics but nearly identical in appearance, finally receiving recognition. This has become a mainstay of the three decades of splitting we have enjoyed since then, from the Western/Clark's Grebe to the Dusky/Sooty Grouse. We can probably look forward to more of these in coming years as well. But these splits have been based on the totality of the species biology, including but not limited to genetic differentiation. The AOU has actually used a conservative approach, and I am sure it will continue to do so. In the future we may well be counting three meadowlarks, three ravens, four screech owls, and two warbling vireos, just as we now count three "solitary" vireos, four sapsuckers, and two sharp-tailed sparrows. But then again, we may not. There's a long way from identifying a difference in the mtDNA genome to determining a species split.


After reading some of the quotes attributed to the study's authors, maybe I'm not so quick to let them off the hook and blame it on the media. Though these quotes are probably over simplified and somewhat out of context, if accurate, they would display a shockingly poor understanding of evolution and speciation. For instance, the quotes suggest unclear understanding of the difference between two populations of one species that have not interbred because of geographic separation, versus two species that do not interbreed because of biologically-based reproductive isolation. And I was downright offended by this quote attributed to Hebert in reference to cryptic species:

They are typically “small brown ground-dwelling shrubbery birds that don’t attract a lot of human attention.”

Um, excuse me? Ornithologists and birders pay no attention to little brown birds? Sorry, gents, but you need to get your heads out of the lab and find out even the tiniest bit about how biology is done in the world of intact organisms before you spout nonsense like this. Library shelves are filled with detailed studies of "“small brown ground-dwelling shrubbery birds."


At 2:10 PM, Blogger John B. said...

Thanks for your thoughts on this. What I found interesting about the press release and press coverage is that there was a lot of hype about the proposed splits and not much attention to the suggested lumps. If, in fact, all of the splits and lumps were accepted by the ornithological community (and that is a big "if"), there would actually be a net reduction in the ABA-area list. So, there was not a whole lot to get excited about, from a listing perspective. But thoughtful analysis is no match for hype!

Plus many articles missed that the point of the exercise was to test barcoding for identifying species, not to decide on splits and lumps.

At 7:22 PM, Anonymous Anonymous said...

FYI, you might wanna check out some of Robert Zink's papers and others that now use the phylogenetic species concept. There's a substantial movement (possibly short-lived, only time will tell) away from the Biological species concept, though like other species concepts, it has its problems, too.

At 6:33 AM, Blogger Bill Pulliam said...

I think the phylogenetic species concept is of limited value in the larger fields of biology and ecology, where we need distinctions that are more fundamentally meaningful. The phylogenetic species concept is a throwback to the ancient morphological species concept: if it looks different, call it a different species. I doubt it will take hold; I deeply hope it does not (in spte of the fact that it would give us birders a lot more tickies).

It is also not in any way a substitute for the biological species concept, as it makes distinctions based on fundamentally different (and far narrower, some might say tunnel-visioned) aspects of biology. If genetic taxonomists want to use this concept, they need to use a different word for the units they define other than "species." That is like redefining a "meter" to now mean what we formerly described as an "ounce."


Post a Comment

<< Home

Site Meter