Predictable Evolution Trumps Randomness of Mutations


Although mutations, the driver of evolution, occur at random, a study of the bacterium Escherichia coli reveals that nature often finds the same solution to the same problem again and again.

Over time, random mutations enable organisms to adapt and diversify, often when geographically separated groups of the same species grow better suited to their local environment and less like members of the other group.

But that’s not the only way that genetic diversity can arise. Researchers have reported cases of cichlid fishpalm trees and finches adapting to different ecological niches and splitting into different species despite living in the same place. In 2008, evolutionary biologist Michael Doebeli of the University of British Columbia (UBC) in Vancouver and colleagues reported that E. coli bacteria can also diversify while sharing a test tube.

In that study, they fed easy-to-digest glucose and a harder-to-stomach acetate to homogeneous populations of the bacteria, and let the bacteria chomp away. E. coli can switch between the two foods, but the team found that in each test tube two groups emerged, specialized in consuming either glucose or acetate. What they did not know was which genetic path each group took to achieve its specialization.

Written By: Lucas Laursen
continue to source article at


  1. The title is disingenuous. We’ve shown that, when the same feature’s evolution is encouraged, the details of how it is effected are the same in each case. Does that mean mutations occur in a nonrandom way, i.e. one which is correlated with their fitness? Absolutely not. Not only is there no conceivable mechanism by which that could be so; not only do we know enough about the quantum mechanics of how mutations occur that we can rule out the idea; but there’s a much simpler explanation of what’s been observed here. Suppose the observed mutations are the “best” (perhaps even only) sequence of mutations to achieve this adaptation, subject to the following criterion: given how likely the relevant mutations are, and given the fact we’re restricted to sequences which yield the required mutation and always effect incremental improvements$, the mean time it would take to happen is minimal. In other words, suppose it’s the sequence through design space to the relevant adaptation which natural selection is the best able to navigate, given the odds of different mutations. But we’ve long known that’s exactly how it works.

    $ Technically, neutral mutations could drift into higher frequency, as could detrimental mutations, although the odds against all but only very slightly detrimental ones doing so is slim to none in a population as large as that used in the experiment. Indeed, these probability commentaries can be translated into a calculation, in the end, of how long on average it’d take to achieve this.

  2. When we look at parallels in evolution arising from different mutations, but driven by the same environmental pressures, the effects of natural selection can be more easily identified. There are many examples of similar structures being evolved through different genetic routes to solve and adapt to the same environmental problems.

Leave a Reply