Taxonomy, phylogenetics, morphology and DNA sequencing.

[DavidCampen]

First let me recapitulate my vague understand of taxonomy and phylogenetics (as applied in botany). Taxonomy is the classification of plants into groups according to physical characteristics of the plant(morphology) and perhaps also population groups and pollinators. A principal reason for doing this is the expectation that the groups delineated by taxonomic classifications will also indicate evolutionary associations (phylogenetics).

So, I am wondering about the current state of DNA sequencing of genes. How expensive is this now? It would seem that DNA sequencing would be much more informative and definitive than morphology in placing plants into taxonomic/phylogenetic groups.

Does anyone have information about this?

[msaar]

Many orchid species are being reassigned based on DNA analysis, most recently and controversially the Oncidinae.

[Andrew]

David,
I'm a molecular biologist involved in medical research so I'll base my answer on what I pay vs what I suspect biologists can afford. Sequencing is fairly cheap at about $5-10/run/sample for contracted Sanger sequencing, which is what most taxonomic sequencing seems to use (less if you have the equipment to do it in house). Invitrogen's new next generation sequencer will do an entire genome for around $1000 (+$150K for the machine).

That being said, cheap is relative. Going off a quick look at the grants available from various botanical/lay societies, there's very little money available for orchid taxonomy. eg the IAPT provides grants up to $1000; the ASBS provides research funding for up to $6000; the AOS provides grants between $500 to $12,000. Realistically, that will buy a botany student a hand lens, calipers, petrol money and some beer to do traditional taxonomic studies but it doesn't go far for molecular research; hence, probably why you get a lot of unrepeated, single sequence phylogenetic studies that fall flat when someone reanalyses the data.

You are correct that DNA sequence data, when produced and analysed properly, is a much more accurate way of measuring evolutionary lineage than morphology. Of course there is still a lot of anti-molecular prejudice amongst the old guard of taxonomy and we horticulturalists seem to be unable to grasp the idea that staying up to date with scientific research is contrary to our demand for taxonomic stability and balk every time a 'taxidiot' dares to change a name.

[tropterrarium]

Taxonomy is not restricted to morphological character, but may, and today often does, include molecular data (DNA sequences).

5-10 bucks per lane (= single sample, single direction; remember DNA has 2 strands, and usually one sequences both strands/directions) sounds about right. Student grants are smaller, but you can get a bunch done. NSF pre-docs are in the 10-15K, full blown NSF/NERC/DFG grants are usually in the 100K to 500K range (at least half for salaries), but you need to be a faculty/curator (or similar) somewhere, and then you provide funding for your students and post-docs. Funding probability with US-NSF is <10%!

Often also, a researcher uses existing data from GenBank, then adds some new ones, and publishes those new finds. So you don't always have to start from scratch.

Re informative, it is a double edged sword. Usually (gross simplification), morphology yields fewer characters, but they tend to be more conserved/stable. DNA sequences yield more basic data, but it is more noisy (depends also on which piece is selected) ITS, 16S, COI, 18S are progressively more conserved, slower evolving. Usually, the particular genetic markers are selected with the goal of the study in mind.

You could also analyze molecules and morphological characters together, though it is rarely done. Most common is mapping of a morphological character onto a molecular phylogeny. True combined analyses are tricky with the more "statistical" molecular analyses such as Likelihood and Bayesian Analysis.

A much bigger problem is how well specimens are identified. Most grad students are no longer trained in classical plant identification (or even just a piece of leaf is obtained, which make ID impossible). Estimates for misIDs of species in GenBank are up to 30%. That is frightening. It is getting a bit better, as people are more aware of the issues, and are depositing vouchers.

DNA is not a panacea. With more emphasis on non-destructive surveys, for instance, photo-IDs are important, and there is no way of extracting DNA from a picture. DNA is one tool in the kit of taxonomists.

For more info see Hillis Moritz and Maybe (spelling not quite sure) "Molecular Systematics" ~2000, a bit dated, but gives a lot of basic info on the topic.

[tropterrarium]

Forgot in cost analysis. Sequencing is only one step. You first need to extract DNA, then PCR the particular gene, clean it. That adds maybe $10-20 per sample. You can PCR multiple genes from one genomic DNA extraction, so it is a bit difficult to give a precise figure. Also depends on what taq you use for PCR. Some are dirt cheap, better ones for difficult genes are more costly.

[DavidCampen]

An online article:
Orchid