5 April 2008

Genomics and plant evolution: blogging on my own peer reviewed research

ResearchBlogging.orgIt's been over a month--that's about seven thousand blog years--since my PLoS ONE paper came out and I'm only just now getting around to giving it a proper blog treatment. In defense of my tardiness, the paper is pretty small-fry in terms of its newsworthiness. It didn't even get a press release (though maybe that's a good thing). In other words, it's still fresh to the blogosphere. Well, that's my excuse at least.

At first I wondered: is this normal? Blogging on one's own peer-reviewed research, that is? But Bora Zivkovic (PLoS ONE Online Community Manager/crazy uncle of the science blogging community) and Liz Allen (PLoS Director of Marketing and Business Development) have made it clear that giving my own paper the BPR3 treatment is not only normal, it is expected.

There are several ways I might approach writing this post. The most obvious is to simply summarise the paper In Plain English. Problem is, I've already done that by writing a news blurb for the Natural History Museum website, and I don't really feel like repeating myself (though of course I will, but only to the extent to which it is necessary to tell my story).

A second approach is to critically analyse the paper. But you can see the problem with that right away: since I wrote it, I've already critically analysed it. Critical analysis necessarily belongs to someone who isn't an author on the paper.

A third approach, and the one I am going to go with, is to tell the whole story of this project: the blood, sweat and occasional tears, not just the part that appears in the paper itself. This will be by far more interesting than a simple recap of the key findings of the paper itself (which, as I said, you can get elsewhere). Moreover, the whole story illuminates the reality of the scientific process in a way that's intelligible to the non-scientist; well, that's my aim anyways.

This is
not my study organism.
About four years ago I was working as a postdoctoral research assistant to Professor Richard Bateman (then Keeper of Botany at the Natural History Museum and president of the Systematics Association) on two very closely related species of British terrestrial orchid, neither of which is pictured at right. I could tell you which species we're really working on but then I might have to shoot you. Okay, it's not that serious--Richard's spoken publicly about this story--but this work hasn't been published yet and I wouldn't want to steal our peer-reviewed thunder on a blog.

So, back to our pair of mystery orchids. Though they are so closely related that their designation as separate species is debatable, the two can nevertheless be easily distinguished by their floral morphology (the size and shape of the various parts of the flower). At the time we began, there was some anecdotal evidence that the two might be discriminated by DNA as well, and thus it was our intention to follow this up with a more focused DNA sequencing project to see if the difference in floral morphology correlates with unique genetic signatures.

Using that knowledge as a foundation and context, we could then perhaps set up an experimental system to pursue the specific genetic underpinnings of the different morphologies. And that's really the exciting part: ferreting out the developmental-genetic mechanisms of evolutionary change (an undertaking called "evo-devo", about which readers of a certain atheist proselytiser's blog will already be aware).

To query the genomes of our two species, I used a common but no less wonderfully clever lab technique called Polymerase Chain Reaction (PCR), coupled with the equally clever Sanger sequencing (which we do using fancy dyes and capillary tubes these days), to obtain short (less than 2000bp) sequence reads from a handful of gene regions that had been previously shown to be useful in reconstructing evolutionary relationships within and among species.

After sequencing more than 10,000bp-worth of these regions from each of four specimens (two from each species), I could not find a single difference between the species. Well, there was one difference, but it didn't sort by species. This kind of evidence doesn't necessarily prove that these aren't different species (while DNA can in some cases identify species it cannot--or at least should not--be used to define them except as a last resort), but it does suggest that the morphological difference that defines these species has evolved relatively recently, that is, so recently that the gene regions I tested have not had a chance to evolve since then. This brings the evo-devo question tantalisingly close: we may in fact have caught this species in the act of splitting in two, in other words, we may be witnessing the origin of species.

But if we couldn't find any genetic variation to distinguish the two species, how were we going to proceed with making our wildest evo-devo dreams come true? There are other methods for detecting genetic variation within and among species (AFLP, microsatellites, etc. for those who care), but I'm not going to go into details on those except to say that they involve a laborious preliminary optimisation phase every time they are going to be applied to a new group of organisms. So I started sniffing around for a different technique, a technique that perhaps could bring the high-throughput technologies of the genome revolution to bear on biodiversity.

That was when I saw a conference abstract presenting a new technique called Diversity Arrays Technology, or DArT. DArT builds on an existing technique called a DNA microarray that is used to detect small genetic differences among people and other organisms for which the full genome sequence is known. The innovation of DArT is that it uses microarrays in a new way so that no prior knowledge of the genome is required. It immediately occurred to me that this opened the possibility that all of biodiversity, not just a handful of so-called 'model organisms', could be examined using the tools of the genome revolution.

DArT was originally used to map naturally occurring (not genetically engineered) traits in economically important crop plants like rice and barley, but hadn't yet been tested on 'wild' plants (or animals). I asked Richard if he thought it was worth pursuing DArT to tackle our orchid speciation question and this was met with great enthusiasm and I was encouraged to write a grant proposal for a small pot of internal funding. I had a couple of project-planning phone calls with Andrzej Killian, inventor of DArT and director of Diversity Arrays Technology Pty Ltd, based in Canberra Australia, and it looked like we were set to give it a shot. Ah, those were the heady days...

But here our story takes a strange twist--the sort of twist that doesn't make it into the peer-reviewed literature. See, unfortunately at this point, there was a shake-up in departmental leadership and the politics of power found their way into the very heart of myelittle project. Due to past and present departmental sensitivities I will not discuss it here except to say that it resulted in a change in both the plants and the biological questions to which DArT would be applied in my proposal. Fortunately, despite this last minute change in line-up, the proposal scored very high marks from the committee and was funded in full.

And so I began the process of testing whether DArT could work successfully within and among wild plant species, as opposed to domesticated, inbred strains of agricultural crops in which DArT had already been proven to work. I applied DArT to two different groups of plants. The green spleenwort Asplenium viride and the moss Garovaglia elegans (left) were chosen for the trial to reflect existing Botany department areas of expertise. Professor Harald Schneider, Dr Johannes Vogel and Dr Stephen Ansell are actively researching the evolution of Asplenium and Dr Angela Newton and Dr Niklas Pedersen the evolution of Garovaglia. They assembled an appropriate selection of specimens and posed specific evolutionary questions that might be addressed by the DArT technique.

I wouldn't be writing this if it hadn't worked. DArT gave us a collection of 1349 new genetically variable regions to work with in order to examine the relationships within and among Asplenium and Garovaglia species. My co-authors' analysis of the DArT data corroborated known relationships and also revealed some new patterns. But the key result was that DArT could be successfully used on non-model, non-domesticated, non-inbred, that is, wild plants.

DArT also proved adept at detecting hybridisation, which is very common in plants, and even mixed-specimen samples. And when I obtained the DNA sequence reads of a smattering of the DArT regions we recovered, I found that quite a few matched known genes in that great big DNA database in the sky, GenBank, and one even matched one of the proposed plant DNA barcoding genes, trnH-psbA (more on DNA barcoding in a previous post).

My hope is that this study will give others the confidence to apply DArT to their research groups, and that this will speed the rate at which patterns and processes in evolution can be understood, and genetic diversity conserved in the wild.

And in case you were wondering whether I still have hopes of applying DArT to our "orchid problem", all I can say now is: watch this space!

For more information about DArT and how it works, visit Diversity Arrays Technology Pty Ltd.

James, K.E., Schneider, H., Ansell, S.W., Evers, M., Robba, L., Uszynski, G., Pedersen, N., Newton, A.E., Russell, S.J., Vogel, J.C., Kilian, A., Michalak, P. (2008). Diversity Arrays Technology (DArT) for Pan-Genomic Evolutionary Studies of Non-Model Organisms. PLoS ONE, 3(2), e1682. DOI: 10.1371/journal.pone.0001682


Mike Haubrich, FCD said...

I am glad that I came back to read this post.

talia said...

As a non-scientist who does not often read peer-review papers but is interested in the life and studies of scientists, I really appreciate your post! Great radio interview too :)

Kevin Zelnio said...

Karen, this is brilliant. I submitted it to Open Lab 2008. You have inspired me to do the same when my papers get published!

while DNA can in some cases identify species it cannot--or at least should not--be used to define them except as a last resort
EXACTLY, but many barcoders don't share your sentiments I think. Variants of the phylogenetic species concept rely on essentially bootstrap values to support new species designations.

Simit said...

Very well written article. I have been aware of DArT marker technology for a short while, and I am very keen on the idea of applying the technology to non-model/wild organisms to address biodiversity questions, rather than economically important organisms.

I hope to see more ecologists taking an interest in DArT markers in the coming years. And it would be good to see this technology being applied to one or two animals too. Thank you.