Update on E. coli long term experiment

Then your specification is completely irrelevant to the domain of biology, where natural selection provides a kind of guiding hand.

A lot of schlock gets read by the public. You don’t have to be a best-selling author to be a successful writer. And you don’t have to be a perfect bacterium to be a successful bacterium. :slight_smile:


Which is why it was no surprise to YEC or ID people.

Can you elaborate on that?

However it had been observed back in the 1990’s and now it has been shown in the lab that it can reliably be induced in E. Coli cultures in a matter of months. Something that can be produced reliably and in a short time is probably no great evolutionary innovation. This is an order of magnitude faster than the LTEE because that did not particularly favour this change.

It would be interesting to see how long it takes E.coli to revert when conditions no longer favour this variety. I wonder if this experiment has been done?

A huge difference is that Lenski wasn’t trying to induce aerobic citrate utilization. The amount of citrate in the medium was, I suspect, significantly lower than that in the studies you are referring to. In any case, it would be very interesting to see what the other mutations have done to allow the altered metabolism. Do you know of any biologists involved with the LTEE or the other citrate utilization studies that claim it is “no great evolutionary innovation”?

I haven’t read all of the LTEE papers, but I don’t know that this has been tried. Agreed, it would be interesting to see what happens in culture medium lacking citrate.

1 Like

If you agree, if you agree that environmental change is the basis for natural selection, and the Lenski experiment verifies this, I welcome your agreement. My problem with BioLogos is that they do not.

As I expect you know, traditional evolutionists, Dawkins et al, say that evolution is random and unguided. This is good for their contention that life is empty and meaningless, but terrible science, because it means that it natural selection is not verifiable and not scientific.

Random natural selection also does not compute. Back in the day when I was young then we discussed evolution it usually ended when someone cited the “monkey example,” if you has a billion monkeys and put them typing on a billion typewriters, for a billion years, they would eventually type War and Peace

I assumed that there was something to this until I started doing the math. If you start squaring 24 for each letter then you will be lucky to write a page over a billion years, much less a chapter.

Then too the whole thing is false. One cannot expect a string of random letters to be rational. That is just not what they are. By definition random strings of letters or numbers are not non-random.

Yes, you might be able to find a some letters that are in the proper order so you can put them all together to make the sentences that you need to make up a paragraph, etc, but this is not random.

Thus the picture that Dawkins paints of evolution is false and irrational. It confirms the atheist understanding of life with bad science. It would seem that BioLogos would want to a 1) Fix the science, and 2) Fix it in a way that would strength our claim of evolutionary creationism. Ecological natural selection that way.

God created humanity by creating our universe in such a way that human being were created by evolution as witnesses to God’s Creation and God’s history. This is the message of the anthropic principle. .

Indeed, they are simply the processes we can observe, in real time, played out over a longer timescale.

We also can’t watch continents jump hundreds of meters, but we can watch them inch (millimetre?) along, and it just so happens that those present-day rates, extrapolated backwards, match up with other lines of evidence that indicate they were once elsewhere -such as the nicely matched up coastlines of Africa and South America.


It would have taken a LOT of materials and labor to capture a new generation every 20 minutes. What they did was start a new culture each day with a 1 to 100 dilution of the previous day’s culture and then let it grow overnight. If you start with 1 and build to 100, the first generation is 2, followed by 4, 8, 16, 32, 64, and 128. So there are about 6.5 generations per culture per day before the media is used up by a 0.1% inoculum. Most of those generations will occur during the log phase of growth early in the culture when there is plenty of food. This is the period of time where generation times are around 20 minutes.


Indeed a LTEE experiment involving 20-minute generations would be effectively impossible without a robotic system. But there is at least one more important scientific reason for the choice of a median generation time: the problem/phenomenon of clonal interference. Lenski has written about this and other really interesting aspects here:


Actually, XLNet text generation is not a random letter generator. It is a language modelling program with several predecessors and an enormous amount of intelligent design put into its development.

… language modeling is a fancy word for the task of predicting the next word in a sentence given all previous words. This seemingly simple task has a surprising amount of depth and the true potential of language modeling started to be unlocked by methods using it as a pretraining method.

Yes, that’s what I said.

E. Coli starts with all the genetic instructions and molecular machinery for citrate metabolism; the bacteria in the LTEE did not have to evolve these during the course of the experiemnt. Normally in E. Coli this is turned off in the presence of oxygen. However the transition to Cit+ in the presence of oxygen is well within what Behe calls the edge of evolution as has been shown by experiments. It was a surprise at the time and widely trumpeted as a major outcome of the LTEE but much less impressive in retrospect.

Hi Chris,

I believe I addressed this idea in another thread, so I beg the indulgence of anyone who recognizes my repetition.

Just as an intelligent community designed the stochastic XLNet language generation model, the God of Abraham, Isaac, and Jacob designed the evolutionary processes that we observe in nature.

We could go a long way with this!

  • Beam search in XLNet is similar to natural selection in biological evolution.
  • Sampling in XLNet is similar to genetic recombination.
  • Online training with backpropagation is similar to mutation and population genetics.
  • Etcetera.

It’s important to pay attention to the level of abstraction. At a lower level, a biologist can identify the dynamics and constraints at work in evolution, and a computer scientist can identify the dynamics and constraints at work in XLNet. At a higher level, a philosopher or theologian can identify the creator of biological evolution, and a historian or devops engineer can identify the designers of XLNet.

You are attempting to reject a low abstraction level theory of biology by making an appeal to a high abstraction level analysis of XLNet. This is an improper argument. If you could instead respect the abstraction levels at work in both domains, you could understand how someone like me can trust both the scientific explanation provided by evolution and the theological explanation provided by Christian faith.

Chris Falter

1 Like

A recent summary by J Wile

To understand what has happened, we need to go back to 2008. In that year, the LTEE showed that even though Escherichia coli normally can’t make use of a chemical called citrate when oxygen is present, one of the their populations developed that ability after 31,500 generations of existence.1 As a result, it was dubbed the “citrate plus” population. How did that happen? At the time, no one knew. However, evolutionists thought it was the result of some rare event or combination of events, exactly the kind upon which evolution depends. New Scientist put it this way:

By this time, Lenski calculated, enough bacterial cells had lived and died that all simple mutations must already have occurred several times over.

That meant the “citrate-plus” trait must have been something special – either it was a single mutation of an unusually improbable sort, a rare chromosome inversion, say, or else gaining the ability to use citrate required the accumulation of several mutations in sequence.

Lenski himself was bold enough to write:

So the bacteria in this simple flask-world have split into two lineages that coexist by exploiting their common environment in different ways. And one of the lineages makes its living by doing something brand-new, something that its ancestor could not do.

That sounds a lot like the origin of species to me. What do you think?

Not surprisingly, a recent experiment has shown that the evolutionary predictions of Lenski and New Scientist are wrong. At the same time, it demonstrated that the predictions of both intelligent design advocates and creationists were correct.

Not sure I understand. What predictions were correct? It appears to me that evolution took place with the changes that enabled citrate to be used, and in addition it was stated that the E. coli also adapted to the simple life in the flask, which is also evolution. Could you elaborate on how that is intelligent design?

1 Like

Then why wasn’t it able to utilize citrate in an aerobic environment?

That seems to be a bare assertion lacking any evidence. Where are the calculations and science backing this claim?

1 Like

Wile never describes the creationist “prediction” or why it would be any different than that of evolutionary biology. Nor does he offer any support for the notion that the predictions of evolutionary biology were not met by the LTEE.

Finally, if Wile thinks that creationism predicts no speciation, how does that reconcile with AiG’s hypothesis that all species in the cat family descended in the last few thousand years from a single feline pair (that was ostensibly aboard the ark)?

Chris Falter

In the genes of the proto-cat there was a little lion, a little house cat, a little saber tooth, snow leopard, and all other cats, just not dogs. The kittens lost a bunch of genes and devolved to today’s familiar cats. Did you know that lions can mate with tigers and produce ligors? QED.

1 Like

It is interesting that we spend a lot of bandwidth discussing human genetic bottlenecks with the flood, and tend to ignore that it also would have resulted in even more severe bottlenecks in every animal surviving on the ark.
The AIG proposal of these bizarre proto-species is outlandish. It has no genetic basis, no historical basis (no proto-cats in the Bible), no fossil evidence, and most important from the AIG view, no Biblical support. I suppose they could claim some transitional fossils are young and represent the cat kind, but that would seemingly admit that they were transitional fossils as well. The AIG ark truly is full of imaginary animals.


One could also add no time, seeing as house cats, cheetahs, lions, and leopards are all represented in early Egyptian art dating back to before the Bishop Ussher flood date in the most accepted chronologies, and even in the untenable AiG ANE chronology only allows a few hundred years for the cat radiation.


Perhaps I’m being overly critical, but I find this trend with several YEC explanations for currently observed phenomena, such as astronomical objects being billions of light-years distant. “Our hypothesis is that xxxx happened, and although there is no scientific or Biblical support for this, it would explain our current observations.”

1 Like

Yeah, I’m pretty sure when Moses wrote Genesis 1, he had in mind that a lion will give birth to a lion, a tiger will give birth to a tiger, a house cat will give birth to a house cat, etc. And that’s entirely consistent with real evolution. The hyperevolution idea seems decidedly anti-biblical.


“Let your conversation be always full of grace, seasoned with salt, so that you may know how to answer everyone.” -Colossians 4:6

This is a place for gracious dialogue about science and faith. Please read our FAQ/Guidelines before posting.