New Science Of Metagenomics to Transform Modern Microbiology?

ScienceDaily has a look at the emerging field of metagenomics that watches the DNA of whole communities of microbes to better understand the microbial world. "Metagenomics studies begin by extracting DNA from all the microbes living in a particular environmental sample; there could be thousands or even millions of organisms in one sample. The extracted genetic material consists of millions of random fragments of DNA that can be cloned into a form capable of being maintained in laboratory bacteria. These bacteria are used to create a "library" that includes the genomes of all the microbes found in a habitat, the natural environment of the organisms. Although the genomes are fragmented, new DNA sequencing technology and more powerful computers are allowing scientists to begin making sense of these metagenomic jigsaw puzzles. They can examine gene sequences from thousands of previously unknown microorganisms, or induce the bacteria to express proteins that are screened for capabilities such as vitamin production or antibiotic resistance."

Comments

how do you...

[jhfry]

extract dna from millions of microbes?

I always thought that DNA extraction was a manual process... or at least it required a significant amount of manpower to get.

Re:how do you...

[Dr. Eggman]

Even if it is manual, there's nothing that says each piece of DNA has to be extracted one at a time. It could be done enmass by taking 'millions of microbes,' shredding the cells and running them through some sort of filter or enzyme that removes the cellular material and leaves the DNA as atleast some fragmented wholes.

Re:how do you...

[zippthorne]

I am not a biologist, but I was under the impression that current gene sequencing techniques involved taking a rather large sample, mashing it up to break the cell walls and release the DNA molecule from the nucleus, introducing enzymes to further break the strands into smaller, manageable lengths, {magic, wherin those shorter molecules are actually translated to bits on a computer somewhere}, then pattern matching to splice the pieces back together digitally thereafter.

In theory, this would sequence everything in a sample, but depending on similarities, could be pretty error prone.

Re:how do you...

[jhfry]

Ok... I get it now... I just learned more about DNA extraction in the last few minutes as I researched this on my own then I ever thought I would know!

Gotta love curiosity. We really need an educational system that fosters curiosity and research above all else... it makes learning so much more fun.

Thanks for the reply!

Re:how do you...

[Vexorg_q]

Extracting DNA is actually a somewhat trivial process, easily done at home with common household products.

Its just a matter of breaking down the cell membranes- which are essentially fatty acids with proteins in them, easily dissolved by detergents. Next, separating the DNA from the rest of the cytoplasm, by putting the extract in alcohol.

A more challenging aspect is preparing the extracted DNA for analysis; you have to clone and amplify the DNA in order to make a DNA library, that becomes more expensive. Nevertheless, processes and enzymes to make DNA libraries have been around for at least 30 years.

The hardest, and most interesting part of this new field of Metagenomics, is making sense of all the data, and observing changes in the DNA as the fauna react to the environment around them. But actually getting the DNA out, thats high school biology.

Re:how do you...

[drooling-dog]

I always thought that DNA extraction was a manual process... or at least it required a significant amount of manpower to get. Nope. That's pretty much been automated in the laboratory, as has much of the analysis. A lot easier than sorting and isolating individual critters, anyway...

Re:how do you...

[goombah99]

Meta genomics is usually applied to unculturable communities. As such it can only be done when the source is so abundant that one can get enough DNA to be able to sequence it.

The best this can do is tell you what genes are present in abundance. Often you may also need primers for that gene so you have to guess a portion of it before you go looking for it. Thus one has some blind spots but these are no worse perhaps than the simple reality that one must always miss some of the low concentration genomes. The presumption is that higher concentration genomes are the most important. That's debatable. If a martian sampled our planet he'd conclude we are irrelevant, and probably that nothing but the top layer of sea water was relevant, given the profile of DNA concentration. Maybe he's right, but I think he'd be missing out on using this to explain a lot of phenomena on earth. How would this explain for example high rise building, deforestation, or changes in the atmosphere, let alone nuclear explosions. For those you might need to sequence us.

Another problem with this kind of analysis is that while it tells you what is there it does not tell you how the genes interact. For that you need to measure things under varying condittions where relative abundances shift. E.g. finding conditions where nominally the same populations exist--highly coupled envirnonments in equilibrium--where there are different stresses and opportunities. Perhaps the best example of this is depth profiles in sea water. However, obtaining enough degrees of freedom in the experimental conditions, so that one can correlate DNA presence patterns is rough. These self-simmmilar variations can be factored out only under assumptions that need to be justified. Typically Linear factors are assumed and that's almost certainly not true. It certainly would be false in any situation involving either negative feedback or saturation effects. getting enough sample points of entire meta genomes is thus the limit. It's pretty heroic to do even one. And of course one replicate is not enough since one can't distinguish noise from variations one is seeking. So it's all very hard.

Thus it's sort of a race which will prove more powerful. Reductive decomposition of a population one species at a time or a discovery based meta genomic analysis.

the simple answer is we need to do both. When it works reduction is far more conclusive about interactions. But there's likely some aspects of community life that dont reside in any one geneome but are traits that float around between different "owners". Likewise, most environments like ground soils have proven to be unculturable so one is sort of stuck with metagenomics or nothing.

Re:how do you...

[DragonWriter]

Even if it is manual, there's nothing that says each piece of DNA has to be extracted one at a time. It could be done enmass by taking 'millions of microbes,' shredding the cells and running them through some sort of filter or enzyme that removes the cellular material and leaves the DNA as atleast some fragmented wholes.

Which is, as I understand it (my wife does DNA extraction as part of her job) how DNA extraction is done, anyhow, whether its from a single multicellular organism or a mass of (normally, relatively homogenous, from a common colony) microorganisms.

Where metagenomics is different is in taking samples not from a cultured colony, but from an environmental sample, and incorporate various laboratory and computational techniques to enable analyzing and isolating those of particular species within the sample without culturing them, which makes it more practical to study organisms that are difficult or impossible to culture.

New paradigm

[drooling-dog]

This is really a new paradigm for microbial ecology. Instead of worrying about how thousands of different species (most of them unknown) are interacting with each other, you can now think about what genomic and proteomic resources are present in a habitat. Think of the organisms themselves as just the bags that contain what you're really interested in looking at, and suddenly a lot of insights and high-throughput techniques open up to you.

Source

[Red Flayer]

The article was taken from a National Academies press release. Here's the full report, parts of which (maybe the whole thing? I didn't check) can be previewed as a pdf if you don't want to purchase the book.

Oh, and here's a brief (4-page summary) of the report.

Woulda been nice to have the source info in the summary...

Related Resource

[Anonymous Coward]

The Craig Venter Institute's Global Ocean Sampling Expedition has been collecting Metagenomic samples for the past couple of years. Among other things the expedition has doubled the number of putative proteins. An excellent video from the expedition is available at http://plos.cnpg.com/lsca/webinar/venter/20070306/ index.html and a set of recently published papers from the expedition are available for free at http://collections.plos.org/plosbiology/gos-2007.p hp

A website hosting the data from the expedition catered towards use on metagenomic samples has been developed by the Venter institute and is available at http://camera.calit2.net/

Re:any biologists in the room..ermm...slashdot?

[Manwe's Herald]

I think you understand the steps in reverse order.

1. a mutation happen randomly in sperm or egg.
2. a new queen is born from this mutated reproductive cell.
3. mutation is positive (e.g. the slave from this queen are more efficient)
4. the queen give birth to more new queen than one with less efficient slaves

Let's look at it in another way.
Infertile workers are like our cells. You can have one white cells which is resistant to HIV, but this mutation won't be passed to your offspring. But maybe one of your sperm (the one) will have the mutation which give resistance to HIV. So your son will be resistant to HIV. He will have more chance of surviving and also to reproduce.

Evolution is a very slow process. Mutations happen relatively frequently. Positive mutations are rare. It's also rare that it is present in germinal cells and rare that an offspring is produced from it. The advantage gained is not always enough to allow it to be propagated. it's very rare that all the conditions are met. That's why it's so slow process.

Bacteria are a lot more efficient at propagating mutations than us.

And concerning bees, the workers have the same genome as the queen but some genes are only activated by the royal jelly given only to future queens.

Re:any biologists in the room..ermm...slashdot?

[DragonWriter]

You see, that's where I don't get it: worker ants can be higly specialised, for instance, I've seen ants in the form of (relative) giant sacs which only were useful for storing some sweet excrement other ants gave them. Clearly, a queen having that mutation would die.

The problem with this is that it ignores that not all genes an organism has are necessarily expressed and that, particularly, the expression of genes may be triggered (or suppressed) by environmental conditions or by the presence or absence of other genes. Colony insects have evolved to very effectively exploit this.

So, in short, a queen with the mutation would not die, because the mutation would be dormant in a queen.

Are you implying that all the mutations of all the different workers are in the genes of the queen (and her eggs), but she herself expresses none of the phenology of it?

Yes, all the genes of the various castes are present in either or both the fertile male and fertile female individuals. They clearly aren't all expressed in the fertile individuals, nor all expressed in any of the various infertile castes; which are expressed and which suppressed depends on the environmental triggers to which the inviduals are exposed (mostly, feeding in the larval stage) which create the castes, and on genetic factors (such as those between males and females).

Marketing aside, keep it in perspective.

[cerebis]

I'm not sure whether the above post should be marked "astroturfing" but it sure reads a little too positive.

454's sequencing technology is a welcomed addition to existing technologies, but don't believe the hype, particularly when the person talking has stock options.

The analysis of genomic sequencing data (metagenomics or otherwise) is highly benefited by large contiguous pieces or ideally whole contiguous genomes. Related to this and more fundemental is the fact that the shorter the pieces of DNA spat out by a machine the harder the problem of assembling them into larger contiguous chunks. This is due in part to the combinatorics of an alphabet made up of only 4 symbols but mainly the fact that genomic DNA contains many repeat structures even in lower organisms.

Without going into detail, it suffices to say that the longer the pieces (or "reads") produced by a sequencing machine, the easier the problem. Add to this the realities of sequencing errors and throw in metagenomics where you may have many organisms with almost the same genome, the problem gets quite hard.

Currently the large sequencing facilties that use 454 machines use them to complement their existing machines which produce 3-10 times longer reads (depending on who's talking). There are in fact papers investigating the ideal ratio of reads produced by new and old technologies.

Another factor to keep in mind is that, although the new high-throughput technologies (454 is the first to market, but not the only player) hold alot of promise, a large part of their appeal was going to be an enormous cost reduction. The problem is, so far that part of the equation hasn't met expectation. They are quite costly to run due to the cost of consumables and those prices are set by the manufacturer.

Re:any biologists in the room..ermm...slashdot?

[Daniel Dvorkin]

only one thing I have difficulty understanding with evolutionism (which I am a strong proponent of)

You're lying. Seriously. Only creationists use phrases like "evolutionism" and "darwinistic evolution."

Sad tendency

[mapkinase]

The advent of metagenomics is accentuating a sad trend in science: less lab work, more computers. Do not get me wrong, I feed my kids from the computer desk and I have never touched an "Eppendorf" or "Pipetteman" (not sure about spelling). In the race for grants we are chasing aggrandisement of the projects we are applying to NSF and DOE. More computers, more modeling instead of experimenting.

I have been reading scientific literature for almost 25 years and the tendency is clear: the results of "computer experemints" (read, modeling) are trusted more and more without any experimental verification. The procentage of sequences in GenBank and Refseq which function is determined only by homology to existing proteins grows. That means we are guessing the function of new proteins by comparing them to the proteins which function we also guessed by comparing to earlier proteins, etc...

Number of protein folds is limited: 700, 1000, 30000, does not matter: it is limited, but it does not mean the functions are limited in the same way. How on earth are we going to find out the function of completely new protein that have not enough similarity to anything in the database? We cannot do it on computers.

And obviously we do not have resources to research experimentally 1.5M genes in Refseq. So instead of blindly pumping more and more raw data into our RAID arrays, we need to be more focused on researching the genes, proteins, pathways that have a direct impact on medicine. You know, "stuff that matters".