Ecological Significance of Microdiversity: Identical 16S rRNA Gene
Sequences Can Be Found in Bacteria with Highly Divergent
Genomes and Ecophysiologies

Elke Jaspers and [Jo¨rg] Overmann

Applied and Environmental Microbiology, Aug. 2004, p. 4831–4839

Major Issue:
The conclusion of this paper can be read in the title. It really isn't all that complicated, nor is it unexpected to any evolutionary biologist.

Basically, the 16S rRNA doesn't change very fast, and changes are discrete and rather random, and some changes demand compensatory changes. So, all this being said, there is going to come a time when the rRNA fakes right while the rest of the genome is running left, and you'll end up with an incongruity in the phylogenies at a small scale, maybe at the level of genera or something. This could of course be massaged out of the data if we knew how to weight various parts of the 16S sequence, but we use this data precisely because we don't know how, and we daren't use other housekeeping genes for broad phylogenies.

In essence, we are depending on something that indicates gross phylogenetic position to do too fine a scale phylogenies because we are ignorant of the details within these various clades.

On top of this, we tend to use phylogeny as a proxy for all sorts of taxonomy, including ecological functional taxonomy. Everyone knows that this is purely a convenient, slapdash way to do things. After all, just thinking about ecology in general, plants, bacteria and fungi can all serve as primary producers. Protozoa and bacteria and fungi can serve as detritovores. Fungi, protozoa and plants can be predators at the micro level. It goes on. Phylogeneny has little to say about broad ecological roles. On the opposite end, we know that many pathogens are closely related to symbiotes, but have a very different lifestyle because of a small cluster of genes. Similarly, apes and humans function very differently in the ecology. Even black and white rhinos are very different, because one eats grasses, the other shrubs.

So, all in all, this conclusion that the 16S doesn't represent phylogeny frequently, and further that it doesn't represent ecology, should not surprise us. What is sad is that many people pretend as if it does. They get a whole bunch of 16S sequences out of an environmental sample and make bold statements about how many sulfer reducers and how many this and that are in the environmental sample.

Ignorng the paper's somewhat controversial use of 'ecotype' language, the result that they have put forward should thus is nothing more than revealing the emporer's new clothes in certain aspects of molecular ecology. While everyone is busy trumpeting PCR bias and DNA recovery biases, the very use of 16S rRNA to characterize communities is a dreadfully inadequate shorthand, and even if done perfectly says very little without systematically confirming the identity of the various 16S rRNA bearing strains, making sure there is a 1-to-1 correlation between the sequences and the strains (or species) and their ecological functions.

In the current environment, this sort of research seems hardly practical - which leads us back to an impasse. 16S rRNA is great at telling us something about the bugs, but not good at telling us much about them. A fingerprint of a community is likely to help us highlight when a major change occured, simply because major changes that compensate for other ones in the fingerprint should be very rare, but it tells us relatively little about who disappeared, or what replaced them.

For anyone relying too much on 16S rRNA in microbial ecology, this should be a scary paper. It points to the continued need for specialists at he organismal level in the various clades, people who can use other genes, specially functional enzymes under selective pressures, to identify what is going on in an environment.

Plankton Diversity in the Bay of Fundy as Measured by
Morphological and Molecular Methods

M.C. Savin J.L. Martin, M. LeGresley, M. Giewat and J. Rooney-Varga

Microbial Ecology Volume 48, 51–65 (2004)


Basic Problem:
In days of Yore, as they say, you could identify plants and animals by what they looked like. Morphology.
Then came the hordes of rampaging microbiologists, who realized that classifying all bacteria into "rod shaped (long and short), round (clustered this way and that), comma shaped, and spiral" wasn't going to cut it. They started using biochemical tests instead: ferments glucose, produce gas, produces acid from lactose, etc. These biochemical tests often had the virtue of being related to industrial applications or medical purposes (hemolytic, etc). These were called Phenotypic tests. More sophisticated classifications were done with serotyping, which relied on the otherwise difficult to observe phenotypes related to cell surface antigens.

Well, the microbiologists pressed on and on, but meanwhile, algae and plankton had not really ever entered into their sphere of influence, and had remained the domain of botanists. Botanists, used to a wealth of informative morphological characters, just picked up the microscope and kept looking for the shapes of leaves, etc. Luckily for them, algae and diatoms and such had more morphology than bacteria. So, their method was more or less successful at creating some classification system.

Well, this has led to a crisis for plankton taxonomy. With plants and animals, some traits have been deceptive, but many morphological characters have been quite informaitive relative to phylogeny. Microbiologists' tests, as it turns out, bore at least some congruence to the presence and absence of key housekeeping genes and pathways, and led somewhat naturally to the current species conceptions that rely on phylogeny (medically relevant characteristics have had a more difficult transition). While discussing and debating bacterial species could take a several volume monograph, suffice it to say that at least there is some value in continuing to talk about the biochemical tests even if you prefer MLST-based taxonomy.

This paper attempts to discern whether there is any hope at all for the existing morphological taxonomy of plankton to have relevance in the modern ecological and phylogenetic taxonomic frameworks.

The Methods:
The choice of the bay of Fundy is naturally arbitrary. I hope it was convenient, because as a vacation spot it seems lousy. The plankton of Guam, on the other hand, should be fascinating...
Anyway, the morphological examination was done in the standard way, I hope (I can't really tell, not being an expert in plankton). The need to use an SEM to identify your taxa indicates to me that morphology is probably a big pain with these plankton, and that they may not be losing much by having to switch to molecular methods, unlike students of some other fields with more convenient but now outdated taxonomic strategies.
They chose to do a sort of ribotyping (DGGE fingerprint), using the 18S (like the bacterial 16S, a slow but steady, broad-spectrum gene across eukaryotes). They also wanted to do sequencing, but it was apparently difficult and they had to resort to cloning to make it happen. The DGGE typing also allowed them to do a 'quick' sort of community analysis. They used bead beating technology to minimize genomic DNA lysis bias, a good idea.

Their results seem solid if problematic. For one, most of their sequences landed in that no-mans-land of parts of the tree where there are other sequences from unidentified, uncultured, uncharacterized environmental isolates.
Worse, "Very few of the organisms identified by morphology were also identified in phylogenetic analyses." In some sense, this is inexcusable - they ought to have been able to try something like aggregating several cells together, or culturing them, to do 18S. However, that would have been a substantial effort, maybe worthy of another paper altogether. At least it should be possible.
The reverse, finding morphologies for rRNA sequences, is probably impossible. And given the extensive discussion of molecular method's biases, which anyone should keep in mind doing molecular ecology, it is clear that quantitative work with environmental samples like these will be difficult to validate.

It hamstrings the paper, though, that they weren't able to find clones for the morphologies and morphologies for the clones. It is known that the two taxonomies aren't congruent, but one would have expected to at least have sequences for the major morphological types.

All in all, this paper does answer one question: studies done the different ways won't be compatible at all. This is a pretty frustrating result, but it isn't entirely unexpected. Direct counts and molecular methods are frequently at odds.

The question of how to improve this state of affairs. Something has to be done to improve our confidence in molecular methods, make sense of the vast and unweildy data a molecular ecology survey returns, and correlate this all to the morphologies. How to do this... especially for plankton... I'm not at all sure.