Single-locus phylogeography studies are now “unlikely” to be published in Molecular Ecology, the most important journal in phylogeography and molecular ecology, based on a new policy (other journals seem to be going this way, too… like Molecular Phylogenetics and Evolution). I ask what this means for authors/studies with single locus datasets, and whether this will contribute to a “file-drawer problem” in phylogeography and phylogenetic systematics.
The “New” guidelines for Mol Ecol and existing guidelines for MPE say NO! to single-locus phylogeography studies
A few weeks ago while considering where to send a new phylogeography manuscript, I checked the Molecular Ecology website and, to my surprise, I found no information for submission guidelines. I took this to signify that the Author Guidelines section of website for the most important journal to the field of phylogeography was under construction, possibly because new guidelines were being drawn up by the editorial board, administration, etc. Later, I began to check back daily to see whether new guidelines had been posted.
Today, I found the Author Guidelines for Molecular Ecology, newly published online. I was shocked to find that Mol Ecol seems to be taking steps towards curbing single-locus analyses/paper submissions. Admittedly, this is just my interpretation but I find it a bit striking that the journal that has probably published more single-locus phylogeography studies than any other journal out there is now turning its back on them. While Mol Ecol has not said that it will not accept single-locus manuscripts anymore by any means, their guidelines website section entitled “Policy Regarding Number of Loci, Populations and Individuals” reads,
"Authors of phylogeographic studies should base their inferences on multiple loci: our editors and reviewers often question the reliability of inferences based on a single locus and such manuscripts are unlikely to be sent out for review."
This policy leaves open the possibility that single-locus phylogeography studies can and possibly will still be published in Molecular Ecology, but this will only happen assuming that, other than a successful review, two criteria are met: (i) sampling strategies should be designed to best address the questions motivating the study, and (ii) marker choices should be designed to best address the question motivating the study (these are nearly direct quotations)… But, to be sure, the key operative word in the quoted statement is unlikely. Deemed unlikely to be of interest, unlikely to be reviewed, unlikely to be considered–but when single locus phylogeography studies also advance science??? Huh?
Apparently, Mol Ecol is following in the footsteps of the journal Molecular Phylogenetics and Evolution, another prominent peer-reviewed journal in evolutionary biology that a couple of years ago (I think) said they weren’t really interested in single-locus phylogeography studies either, so they nixed them in their guidelines. To verify this was still the case, I checked the current online MPE guidelines for submission and found the statements that
“Phylogeographic studies will be considered for publication if they offer EXCEPTIONAL theoretical or empirical advances.”
“Papers based on few taxa, single molecular markers, or that use codon-based methods to test for positive Darwinian selection, but in the absence of experimental evidence that allegedly selected amino acid changes cause an adaptive phenotypic effect, will not be accepted.”
Although the long and boring clause about adaptive phenotypic effects is very distracting, the key point is that this is basically saying that the editorial board of MPE reserves the right to boot into the street up front any study that has “few” samples (subjective interpretation as a basis, or hard criteria?) or is based on single markers (e.g., mitochondrial DNA). But this makes one wonder whether the important connection that can be inferred to exist between these statements isn’t that single-locus phylogeography studies cannot by fiat make exceptional theoretical or empirical advances. Maybe.
Does this reflect an attitude that single-locus phylogeography studies are de-valued as science?
Now, I have several questions. I immediately question whether these trends in Mol Ecol and MPE, and especially the new guidelines in Mol Ecol, are even news at all. If not, then these can be deemd “not a departure from business as usual,” or not worth nary a mention. And I realize that there is nothing surprising about a journal saying it expects the study designs of submitted reports to be appropriate. This is obvious. Also, at the present time in our field while many are surfing or learning to surf on the waves of the statistical phylogeography and genomics revolutions, I admit it’s kind of cool for journals like Mol Ecol and MPE to step up and say that they’ll only take multilocus studies seriously, since single-locus studies have a variety of caveats and limitations. “Yay” for scientific rigor! And we can all see the benefit of what they’re trying to do.
But do these policies reflect a growing attitude among researchers that single-locus phylogeography studies or phylogenetic studies are no longer of use to our science, i.e. worthy of publication? Is single-locus phylogeography now “de-valued” science? The Mol Ecol statement seems to imply this to be the case–such studies are “unlikely” to even be reviewed, whoever knows _how _unlikely. If so, then I think some people will take an issue with this. I mean, using data to test really cool hypotheses while acknowledging the limitations of those data is what we do, and this is no different for multilocus or single-locus phylogeography studies. Moreover, we have to take into consideration that not everyone in the world has access to funding for multilocus studies; not everyone in the world has access to next-generation sequencing techniques; and, for many study taxa and systems where workers make painstaking boots-on-the-ground effort to get samples for phylogeographic studies or population genetics studies, there are simply not enough variable nuclear markers developed that are suitable for generating multilocus datasets that are appropriate for addressing the kinds of questions that phylogeographers are keenly interested in. So, life’s not fair. But it might be more important to ask,
Does this kind of thinking contribute to a “file drawer problem” in phylogeography?
We would all do well to sit back and think about the potential issues with the above kind of thinking that demeans single-locus phylogeography studies. At best, since phylogeography publications in Mol Ecol and MPE essentially have a minimum threshold for sampling (at least, on paper they say they do), we can expect there to be a potential increase in the average quality of phylogeography studies in these journals (I assume this was among the primary intentions of the journal editors/leadership, in both cases). Also, other journals can expect an influx of phylogeography studies that Mol Ecol may not want. So, the news is not “all bad.“ Authors may take their well-written mtDNA phylogeography papers to open access journals (paying thousands of dollars to publish sound science in PLoS ONE or BMC Evol Biol, etc.), or to journals like the Journal of Biogeography that publish excellent phylogeography studies, regardless of the number of loci.
At worst, however, maybe we should all worry just a little bit about this MPE-Mol Ecol trend, and the attitudes it represents. Perhaps the most practical issue is publication bias, and with this in mind I ask, “Will this attitude against single-locus studies contribute to a ‘file-drawer problem’ in phylogeography and phylogenetics?”
To explain, when there is “publication bias,” the probability that a study is published depends on some external factor other than its core scientific merit, including its patterns of statistical significance, the quality of the data, or some other factor. Thus, while a larger pool of studies (and knowledge) exists, the literature will represent a biased subset of all relevant studies in a research area. The classic “file drawer problem” is created when some of the studies in the broader pool of studies that could potentially be published are thus set aside (e.g. in a file drawer, or filing cabinet) not to be published or to be published at a later point in time. This is problematic for a variety of reasons. Chief among them, it means that there will be a difference in the content of what is reported in the literature versus what is actually known by the scientific community, and this messes up meta-analyses among other things. Publication bias is also bad because it wastes money (virtually all resources are funded, with the goal being discovering and publication of new theories, hypotheses tests, and observations), and of course it wastes time.
If my sense of the above attitude from the field and from the guidelines of Molecular Ecology and Molecular Phylogenetics and Evolution is accurate, then this may mean that some authors feeling less confident about their chances of publishing their phylogeographic work in reputable journals like Mol Ecol may now sit on their datasets and not publish them until data for more loci become available (or never at all). And, if this happens–and I’m sure it is happening and will be happening in the future–then this will create (more of) a “file-drawer problem” for phylogeography, and also for phylogenetics. So, if you’re doing any meta-analyses of phylogeographic patterns, beware. Also beware if you are trying to interpret available phylogeographic patterns in your study area. If it is a study system of great value that other molecular ecologists with phylogeographic tendencies are also working in, there is a good chance they have unpublished data that they are sitting on. So, you should do the right thing and request that they share their data or results with you.
I would love to hear what the molecular ecology community thinks about this.
A caveat: more multilocus, genomic and statistical directions represent great progress in phylogeography and phylogenetics
It would be remiss not to point out that I wholly support the direction of our field towards expanding multilocus and genomic approaches, as well as adopting more statistically saavy and computational modes of thinking to frame and test hypotheses in phylogeography and phylogenetics. In particular, I support multilocus statistical phylogeography. I also support integrating phylogeography with other kinds of biological and geospatial data. And I am actively trying to do this in my own research. The purpose of this post is to think about a specific question relevant to sampling and the value of studies that do not meet certain sampling policies for journals. And I have legitimate concerns about several related issues. However, I do not believe that single-locus phylogeography or phylogenetic studies represent a best solution, nor do traditional gene tree or “pattern-matching” approaches. And of course I am not advocating DNA barcoding or other completely single-locus approaches. I’m also not making any claims here that single-locus studies provide the best estimates of population parameters or are able to correctly infer the species/population tree of interest in a phylogenetic or phylogeographic study. It’s just that when people want to outlaw studies that test hypotheses and do science because they use a single locus, this seems reckless, particularly if those studies acknowledge the limitations of their data and present a cautious interpretation of their results. More later.