Natural selection observed in a lab

June 10, 2008 4:36 PM

In the 1980s, Richard Lenski hypothesized that his research team should be able to watch random mutations and natural selection taking place in a lab by observing a bacteria population over many generations. In 1988, beginning with a single bacterium, he started several replicate colonies. Recently, after 33,127 generations, his team has observed natural selection.
posted by Tehanu (55 comments total)

Pffft. Call me when they evolve EYEBALLS.
posted by GuyZero at 4:39 PM on June 10, 2008

This is cool.
posted by CitrusFreak12 at 4:45 PM on June 10, 2008

First, wow, great link.

Second, I love the YouTube video and how they stack the petri dishes to make it look like a Disney-esque castle.

Third, this is pretty amazing. I swear I read of a similar experiment before, but since my google-fu is failing me, I will assume that I am wrong and that this is really especially amazing.
posted by Joey Michaels at 4:45 PM on June 10, 2008

Good stuff and good science. I really liked the fact that they could run back in time and see when the key mutation occurred.

Sure, this won't stop the anti-evolutionists - after all, you don't see a new species appearing and they've long ago conceded that you can breed creatures for certain characteristics - but it's one more nail in their cross.
posted by lupus_yonderboy at 4:46 PM on June 10, 2008

I'd prefer to think its one more nail out of their cross...
posted by Joey Michaels at 4:51 PM on June 10, 2008 [6 favorites]

This is really cool, but I am curious how much freezing the old bacteria damages them.
posted by thrako at 4:58 PM on June 10, 2008

There's a follow up entry on the blog answering some reader questions about the original summary of the work. The paper is online now.
posted by i love cheese at 4:59 PM on June 10, 2008 [1 favorite]

Given the amout of pages on the webs nowadays it is a little freaky when someone posts a link on metafilter which you have already visited.

Although I can't help think I followed the link from here in the last week or so? (it is the second one).
posted by Samuel Farrow at 5:00 PM on June 10, 2008

YEAH WELL GIVE ME A CALL WHEN A CAT EVOLVES INTO A DOG
posted by Sticherbeast at 5:04 PM on June 10, 2008 [5 favorites]

No double popped up when I previewed or checked tags. Hunting around does show a link to the blog post as the first comment in this thread though. Was maybe that it, Samuel Farrow?

If it is a double, well, it's a nice night for a deletion, and apologies.
posted by Tehanu at 5:07 PM on June 10, 2008

Fascinating read and outstanding science writing.
posted by bluesky43 at 5:08 PM on June 10, 2008

Lenski started off with a single microbe. It divided a few times into identical clones, from which Lenski started 12 colonies. He kept each of these 12 lines in its own flask. Each day he and his colleagues provided the bacteria with a little glucose, which was gobbled up by the afternoon. The next morning, the scientists took a small sample from each flask and put it in a new one with fresh glucose. And on and on and on, for 20 years and running.

I've been watching too much BG . . .
posted by tachikaze at 5:16 PM on June 10, 2008

This is cool.
posted by CitrusFreak12 at 7:45 PM on June 10

Eponysterical!
posted by ZenMasterThis at 5:17 PM on June 10, 2008

Fascinating read and outstanding science writing.

Just what I was going to say. Thanks for the post!
posted by languagehat at 5:18 PM on June 10, 2008

This is one of the coolest (and most perversely exciting) things I've read in a long time. Thanks.
posted by OmieWise at 5:19 PM on June 10, 2008

Another thank you. Excellent link.
posted by ltracey at 5:21 PM on June 10, 2008

seconding OmieWise
posted by bashos_frog at 5:32 PM on June 10, 2008

Well its only microevolution, of course that exists.

But seriously, this makes me very happy. I am sure IDers will come up with some way to explain this away, but the more evidence we keep compiling, the more ridiculous the will appear.
posted by [insert clever name here] at 5:32 PM on June 10, 2008

Very awesome and I've often wanted to do some kind of similar amateur with algae.

That said, this convinces 0.00000 IDers. Their entire argument rests on the fact that "you never see dogs turning into cats AMIRITE??" Some pointy headed einstein in an ivory tower with some back-tear-ee-uh that prolly don't even really exist....pff.

Now, do it with beer-producing yeast and you might have something. That's evolution that is relevant to, and appreciable by, Joe Sixpack
posted by DU at 5:36 PM on June 10, 2008

YEAH WELL GIVE ME A CALL WHEN A CAT EVOLVES INTO A DOG

1) D'oh, but
2) I rest my case.
posted by DU at 5:37 PM on June 10, 2008

Well, the ID response could be simple.

God is just messing with the Scientists. He's making those bacteria mutate and just laughing His Divine Behind off.
posted by Joey Michaels at 5:38 PM on June 10, 2008

So wait. This guy is breeding super-bacteria that will eventually naturally select into bacterial warlords and take over the world?!
posted by Saxon Kane at 5:46 PM on June 10, 2008

What do you mean, "take over"? What would change? :)
posted by aeschenkarnos at 5:58 PM on June 10, 2008

If this is the same article making the rounds @ Slashdot, I might add that they have not completely finalized their research; i.e. they have not yet proven that the results weren't totally random.
posted by Dark Messiah at 6:09 PM on June 10, 2008

I can only hope that some of the comments at the end of that post were made with tongue planted firmly in cheek.
posted by JaredSeth at 6:12 PM on June 10, 2008

No double popped up when I previewed or checked tags. Hunting around does show a link to the blog post as the first comment in this thread though.

That's probably it. I posted it as a comment there, though I actually I found the "What is a species?" link in the Loom piece. It definitely deserves its own thread. Good post, Tehanu!
posted by homunculus at 6:15 PM on June 10, 2008

Ahem, God clearly meant for this to happen.
No?
Let's see. This is micro not macro evolution. It proves nothing.
No?
Well, um, God works in mysterious ways?
No?

Crap, well it's a good thing that I'm not a denier of evolution.
posted by oddman at 6:39 PM on June 10, 2008

Citric acid is one of the ingredients of the food they've been feeding their bacteria all this time.

Bacteria other than Ecoli are able to metabolize citric acid, and bacteria of different species exchange genes all the time (as in antibiotic resistance).

If even one living bacterium which could use citric acid happened to be included in the nutrients, that could account for this. It might not even have to have been alive.

Unless the genes which allow the use of citric acid are highly variable and this one turns out to be unique, proving this result is going to be a long row to hoe.
posted by jamjam at 6:55 PM on June 10, 2008

Fascinating read! Thanks, Tehanu!
posted by cowbellemoo at 7:03 PM on June 10, 2008

The bacteria in those flasks has been evolving since 1988--for over 44,000 generations.

I was born in 1988. It's very odd to think of my lifespan in bacteria generations.

Nevertheless, this is cool! I love science.
posted by flibbertigibbet at 7:32 PM on June 10, 2008

jamjam - it would be easy to find out. I assume they tried to extract plasmid from it and didn't pull anything out.

The next step is to sequence the genomes of the different generations. It's gotten cheaper, but that's still going to be a good whack of change and a bit of time. If I was a granting agency, I'd give them the money to do it.
posted by porpoise at 8:11 PM on June 10, 2008

jamjam and porpoise -

Another way of figuring it out would be just to sequence one of the genes responsible for the change, and compare it to earlier copies of the genes in previous generations. Some, but not all, or the changes should be there. It's easy to build a phylogenetic tree with these sequences and show that contamination is an unlikely cause of the mutation.

Dr. Lenski already has the money to sequence these genomes -- it's in the works as we type.
posted by Peter Petridish at 8:43 PM on June 10, 2008

jamjam, they repeated the experiment and duplicated the results. Did they accidentally mix in other bacteria more than once? Much less likely.

lupus_yonderboy, this is a new species -- the evolved trait is one of the things that differentiates this species from others. Also, this is not breeding: the researcher did not choose offspring to mate. And the intermediate mutation is not yet known by the researcher, so he couldn't have bred it in.

This is a dramatic step forward in the study of evolution (and a dramatic new confirmation of the theory). You can only dismiss it by ignoring the facts.
posted by sdodd at 10:48 PM on June 10, 2008

As others have said, this won't disturb creationists at all.

They've long argued that there is what they call "microevolution" causing small changes within species. However, since the Universe was only created six thousand years ago, there's hasn't been enough time for that to create any new species.
posted by TheophileEscargot at 11:12 PM on June 10, 2008

Well, to be accurate they have observed natural selection from the very start of the experiment, as the current evolved strains, when mixed with the original strains, grow much faster. (You mix them at 50:50 ratios, grow them for a few generations and then put them on petri dishes. The evolved ones are white, the ancestral ones are red and you can work out the relative fitnesses by counting the number of white to red colonies.) Here is a graph showing the relative doubling amount over time. After about 20,000 generations the evolved strain is roughly 1.7 times as fit which is a pretty big increase.

This new work is cool, as it shows that rather than "tinkering" with existing traits like most of the other changes seen so far (changes in DNA coiling etc), this is a new trait that has turned up and is a pretty clear candidate for being called a macro-evolutionary step (though it would be nice to see the molecular basis for the change).
posted by scodger at 11:30 PM on June 10, 2008

Fascinating read and outstanding science writing.

It's par for the course at The Loom, Carl Zimmer's blog. It's really a fantastic site.

Seriously, the guy cranks out something this good at least once a week.
posted by delmoi at 12:18 AM on June 11, 2008

lupus_yonderboy, this is a new species -- the evolved trait is one of the things that differentiates this species from others. Also, this is not breeding: the researcher did not choose offspring to mate. And the intermediate mutation is not yet known by the researcher, so he couldn't have bred it in.

However, since the Universe was only created six thousand years ago, there's hasn't been enough time for that to create any new species.

The problem is this argument makes no sense when you're talking about single-celled life forms that don't mate with each other at all. Since a specices is a group of organisms that can mate with each other, "technically" each new bacteria is a new "species" because they will never mate with each other.
posted by delmoi at 12:21 AM on June 11, 2008

Depends on which definition of species you are using.......
posted by scodger at 1:03 AM on June 11, 2008

thrako: As far as I'm aware, freezing the bacteria causes no damage at all. I store E. coli by mixing them 1:1 with glycerol and sticking them in the freezer, and that's the recommended long-term storage procedure from the people selling E. coli libraries so I'm pretty sure it's fine.

This is a really cool bit of research, and an excellent new blog to read. I'm feeling all enthused about science again. Thanks, Tehanu!
posted by penguinliz at 5:59 AM on June 11, 2008

So, delmoi, would you argue that the whiptail lizard found in the southwestern US is not a species? They're all female, they exhibit parthenogenesis (they all lay eggs that hatch as clones of the mother), therefore they never have gene-exchanging sex, and thus by your definition are each technically their own species. Except by any biologically accepted definition, all of them are classified as a species - Aspidoscelis uniparens.

Also, bacteria share genes all the time. Even if they don't do it using plasmids, gene exchange can occur through infection via bacteriophages, or through incorporation of loose bits of DNA from dead bacteria.

Finally this is good news for Rick Lenski. He's a good guy, and this took a lot of work. Rich signed off on the Ecology, Evolutionary Biology and Behavior endorsement on my PhD. He's in charge of the EEBB program at Michigan State now. It's a really interesting, varied and dynamic group. If you're looking for a grad school, and have any interest in evolution, ecology, or behavior, look in to it. Basically any PhD or Masters degree student in the biosciences at Michigan State can apply for EEBB membership as well.
posted by caution live frogs at 6:14 AM on June 11, 2008

Pffft. Call me when they evolve EYEBALLS.

We did, but you weren't home! Now they've evolved jetpacks and are too busy fighting intergalactic crime to talk to the likes or you or me.

Well, to be accurate they have observed natural selection from the very start of the experiment

True. I fumbled on how best to word that.

That said, this convinces 0.00000 IDers. Their entire argument rests on the fact that "you never see dogs turning into cats AMIRITE??" Some pointy headed einstein in an ivory tower with some back-tear-ee-uh that prolly don't even really exist....pff.

Now, do it with beer-producing yeast and you might have something. That's evolution that is relevant to, and appreciable by, Joe Sixpack

The leading people within ID are quite intelligent. It's why ID is so convincing to someone with low scientific literacy-- it was written by people who understand science quite well. Extremely intelligent and extremely wrong people who exploit a lot of logical fallacies to pick fights with science because science keeps finding things that don't fit in their worldview. It's very dangerous to underestimate them. The responses were already out there before I posted this yesterday. Here's Michael Behe's.

As others have said, this won't disturb creationists at all.

It does make a dent, despite their reactions to it. One of their main points nowadays is that a few random mutations plus selection isn't likely enough to ever lead to the development of new and adaptive traits. Which is exactly what happened here, spin it though they will.

If this is the same article making the rounds @ Slashdot, I might add that they have not completely finalized their research; i.e. they have not yet proven that the results weren't totally random.

I don't understand what you mean by "totally random."
posted by Tehanu at 7:05 AM on June 11, 2008

delmoi: The problem is this argument makes no sense when you're talking about single-celled life forms that don't mate with each other at all. Since a specices is a group of organisms that can mate with each other, "technically" each new bacteria is a new "species" because they will never mate with each other.

*ahem* The "species problem" goes back at least as far as Charles Darwin and is a topic of discussion for just about every freshman biology class. That definition has serious problems because 1) it is often very difficult to determine mating patterns in the wild, 2) it doesn't distinguish between genetic, behavioral, or anatomical incompatibility, 3) many populations are not known to have sexual reproduction in natural environments, including plants and fungi, 4) in a few cases, populations are genetically compatible but hybrids have low reproductive fitness and are thus rarely seen in natural populations.

To be blunt, your counter-argument uses a naive definition of "species" that is widely understood by evolutionary biologists to be very problematic, a definition which doesn't apply to microbiology at all.
posted by KirkJobSluder at 8:17 AM on June 11, 2008

One of their main points nowadays is that a few random mutations plus selection isn't likely enough to ever lead to the development of new and adaptive traits.

Which is exactly where this paper comes in. It took 3 separate mutations to create the noted change, which is something Michael Behe says can't happen by natural selection.

With each of the 12 colonies (of Kobol?), a sample was taken and frozen every ~500 generations. One of the colonies was noted to be capable of using the citric acid at around the 33,150th gen. When they looked back through the frozen samples, they noticed a severely-reduced capacity mutation for it around gen 31,000. Re- growing colonies from that point tended to have the strong mutation occur again by the (total) 35,000th gen.

So they went back further, and with each sample they tested, there was a high likelihood of re-developing the trait - until they hit gen 20,000. At that point a neutral mutation had occurred that had the effect of being the stepping-stone for the other two mutations. Colonies taken before that point (and thus before the mutation) showed no further likelihood than the other original 11 colonies of developing the necessary mutation or the final trait.

What this showed is that the history of the colony was important. The fitness of the colony over the others was directly influenced by a neutral mutation - one that had no selective pressures at the time - that occurred at a certain point in their past.
posted by mystyk at 8:37 AM on June 11, 2008 [3 favorites]

Thanks, mystyk; row hoed.
posted by jamjam at 10:11 AM on June 11, 2008

Ow, Behe hurts my head. This is preaching to the choir, but:

The most important concept in a proper understanding of evolution is scale. Evolution occurs over a vast time range and a massive number of trials, not all at once over some drunken weekend. It has taken roughly 3.5 billion years for life to reach this point, and while adaptations appear to be coming faster (as organisms become optimized for change), the key to everything is TIME.

Behe's main point boils down to 'since complex systems require an increasingly long series of random mutations to improve themselves (doubly so if they are developing new traits or processes) the odds of such improvements occuring randomly are virtually zero'.

The reason Behe has given me a headache is that this is true. For every mutation that is required for a new trait to occur the number of generations that will pass before the new appears increases. It's just simple probability. With a genome of X length, and a rate of X/Y mutations per cell division, the odds of a single mutation occuring are roughly 1/Y. When you need two specific mutations the odds are (1/Y)^2 and so forth. There's likely more than one mutation that can accomplish the same effect, but the idea is sound. If any of the mutations in the sequence decrease the fitness of the organism for its current conditions (as is likely), then strains carrying these intermediate mutations will be less populous than the 'neutral' strains, which just adds to the trouble (this doesn't mean they'll die out. In the linked article they mention that the intermediates made up only 0.5% of the cell population, but 0.5% is still a lot of cells).

Where Behe veers off is that he looks at this and says, 'Aha! The odds of this trait occurring are 1 in 100 trillion! Therefore random mutation will never work!' He's missing the scale. Sure the odds are shit, but that matters little when you've got a whole world to play with, and an almost unlimited amount of time to play in. I've heard that our bodies contain 10 trillion human cells, and at least 100 trillion bacteria cells. Given that (under optimum conditions) bacteria can replicate once every 20 minutes, 1 in 100 trillion odds occur on an hourly basis. The odds of a single cell developing into dinosaurs or humans are far worse than 1 in a trillion, but over 3.5 billion years? I'd bet money on it.
posted by Orange Pamplemousse at 8:43 PM on June 11, 2008

Oh, I was just working on a post on this

Here is some more swag:

Carl Zimmer blogs about Lenski's research for the Loom (Scienceblogs)

New York Times article from June '07 on Lenski's work

Intelligent Design advocate Michael Behe chokes out a critique

Article abstract in Proceedings of the National Academy of Sciences (full text requires subscription) (all authors: Zachary D. Blount, Christina Z. Borland, and Richard E. Lenski)
posted by fourcheesemac at 3:42 AM on June 12, 2008

An exciting day -- thanks for the FPP!
posted by fourcheesemac at 3:43 AM on June 12, 2008

And sorry to repeat a couple of the links
posted by fourcheesemac at 3:44 AM on June 12, 2008

a naive definition of "species" that is widely understood by evolutionary biologists to be very problematic, a definition which doesn't apply to microbiology at all.

Here's an interesting bit about the troubles that come up when you try to define bacterial "species." It seems like any definition you choose for the word ends up being "very problematic."
posted by vytae at 1:48 PM on June 16, 2008

Lenski gives Conservapædia a lesson
posted by homunculus at 12:26 PM on June 24, 2008

Whoops, I nearly FPP'd the same link.
posted by GuyZero at 1:36 PM on June 24, 2008

Louisiana passes first antievolution "academic freedom" law
posted by homunculus at 9:44 AM on June 29, 2008

July 1, 1858: Darwin and Wallace Shift the Paradigm
posted by homunculus at 11:35 AM on July 1, 2008

The Darwin Conspiracy
posted by homunculus at 12:36 PM on July 8, 2008

Heck, why not go all the way: every human baby ought to leap right out of the vagina with razor sharp fangs and claws, ready to hunt down and rip the throat out of a gazelle.

That's a universe I kind of want to visit. Gives the whole baby vs. rhino situation a new perspective.
posted by Tehanu at 1:32 PM on July 8, 2008