Sometimes science news feeds are great. They can let one know about papers well outside one’s discipline that are of interest but that would not have be encountered under a normal literature search.
Case in point. There is a story on EurekAlert! about researchers from the University of Illinois using a computer simulation based on evolution to design a more productive photosynthetic leaf. It’s a good illustration of how real-life organisms represent the products of evolutionary trade-offs rather than of optimal design, and of how evolutionary algorithms can result in solutions to complex problems of practical importance. In short, evolution can produce good adaptations, but these are not optimal.
The news release is here:
Researchers successfully simulate photosynthesis and design a better leaf
And the the original paper, which is open access, can be read here:
Zhu, X.-G., de Sturler, E., and Long, S.P. 2007. Optimizing the distribution of resources between enzymes of carbon metabolism can dramatically increase photosynthetic rate: a numerical simulation using an evolutionary algorithm. Plant Physiology 145: 513-526.
Thanks for alerting me to this one. If I really end up writing a book on Darwinian Agriculture, I’ll have to read it. I guess I will anyway.
At first glance the paper presents some problems.
First, it’s using the old-fashioned terminology for photosynthesis. The authors are mostly talking about carbon fixation and carbohydrate metabolism.
Photosynthesis is the production of chemical energy from light energy. The chemical energy is in the form of ATP and NADPH. Those energy currency molecules can be used for all sorts of things like making DNA and protein. They are not restricted to carbon fixation and carbohydrate synthesis.
That’s why we now separate those two phenomena.
The second problem is using mathematical models to reach conclusions about designing a better leaf. This would only be true if you actually demonstrated in vivo that the biochemistry of the leaf worked better.
I’m not convinced that we know enough about the biochemistry to make a good model. And if the model isn’t good then the results won’t be either.
On the other hand, the authors do make the point that normal biochemistry is sloppy and sub-optimal. That’s worth emphasizing.
Absolutely — I am not a proponent of proof-by-simulation. They’ll need to make the plant and see how it does. But the important points that I noted are 1) it is not so difficult to design a series of enzyme doses that might do better than the existing one, because the real version is under other constraints and is not optimal in any one direction, and 2) this can be done effectively using a random mutation and natural selection algorithm more than direct design approaches.
Next step, plant those plants.
Ford —
In the meantime I recommend writing a general article for one of the two new journals that have been launched:
Evolution: Education and Outreach
Evolutionary Applications
There’s a very nice distinction made here:
An obvious question that stems from the research is why plant productivity can be increased so much, Long said. Why haven’t plants already evolved to be as efficient as possible?
“The answer may lie in the fact that evolution selects for survival and fecundity, while we were selecting for increased productivity,†he said. The changes suggested in the model might undermine the survival of a plant living in the wild, he said, “but our analyses suggest they will be viable in the farmer’s field.â€
In artificial evolution we get to write the fitness function.
RBH