The February issue of Nature Reviews Genetics contains the review article “Digital genetics: unraveling the genetic basis of evolution.” Here is a brief discussion of the paper, which is concerned with computer experiments in evolutionary biology.
Adami C. "Digital genetics: unraveling the genetic basis of evolution." Nat Rev Gen. 7 (2) 109-118 (2006).
In silico (computer based) experiments have been on the rise in evolutionary biology for some time now, and Adami’s paper discusses their use, advantages, and disadvantages with special attention to the Avida program. Those who have been around the design debate for a while will recognize the name Avida from a 2004 paper claiming that the program was used to demonstrate the evolvability of irreducibly complex structures (I will not be discussing that claim here).
This paper gives a good rundown on why we are able to get any useful information out of Avida and similar programs. However, the use of computer simulations is a big bright arrow pointing to the inherent informational content of genetics (the clay for potter-evolution). I do not know of anyone denying informational content for DNA. But what I sometimes see is a tendency to pooh-pooh comparisons of DNA with linguistic information content. The use of computer code, whose informatics are well known, to simulate genetics demonstrates that the metaphor of language is indeed pertinent.
Second, a major limitation on current in silico experiments is their inherent reliance on external data. When a programmer introduces selective pressure or other elements of digital experiments they are in fact introducing numbers (e.g. percentages, maximums and minimums). But in the real world, the numbers come from the interactions of the organisms with their environment, and not the other way around. Of course, one of the benefits of these programs is that they allow for “what if” experiments. But these experiments still require constants, and their programming still requires an understanding of biological evolution. Our knowledge of evolution and the constraints involved is far from perfect, and these imperfections will be passed along to any digital experiments even if (and I stress if) the programmers do their best to ensure that the programming and experimental conditions mimic biotic reality.
Digital experiments will, I am sure, continue their rise to prominence in evolutionary biology. They will continue to point research in interesting and new directions. But certainly they should be taken with at least a grain, if not a dash or two, of salt.
