August 18, 2008
Biological Change / The Production of Variability
Most people don't understand the problems of the production of variability in biology (or actually in any complex system), and especially how it relates to Intelligent Design. Here are some links to my current work on the subject:
UPDATE - followup blog entry - applying some of these concepts to spiders
August 13, 2008
Biological Change / The Case for and Against Genetic Entropy
NOTE - this has been updated slightly.
At the ICC, the GENE team introduced a new genetics simulation tool which has an astonishing number of input variables - it seems to be really well thought out. The results of the simulations performed on it show Sanford's "genetic entropy" thesis - that the genetic load that comes with mutation far outweighs any beneficial mutations that may occur.
So the question is, why does evolution go downward? Sanford's conclusions are:
- Near-neutral mutations cannot be removed through selection
- Mutation rate is way too high
- Nonheritable noise (I think this was random death instead of selective death)
- Trait linkage and Muller's ratchet (it was unclear how Muller's ratchet was modelled in the software)
- Fixation of deleterious mutations (note to non-geneticists - fixation means that the mutation isin every member of the population - it does not mean that the mutation has been fixed from being deleterious).
Sanford said that selection breaks down at the 0.001 fitness reduction level. At this point, selection is simply unable to remove the trait from the population. It is also harder to select away recessive genes.
Their software can also simulate a population bottleneck. It showed that this actually leads to a dramatic loss of fitness because of the rapid fixation of genetic damage. So, in the fitness graphs, a genetic bottleneck causes a temporary transition from a downhill slope to a downhill cliff. When the population recovers, it is back on the downhill slope.
Sanford said that this data should cause the following shifts in evolutionary thinking:
- Switching from Primary Axiom (mutation + selection = increasing fitness over time) to Genetic Entropy concepts
- Switching from "forward evolution" to "degeneration"
- Transition from the idea of "creative selection" to focusing almost entirely on "stabilizing selection"
- Thinking about extinction as a past event to thinking about extinction as a future event
Now, personally, I wonder if some of this relies too much on Darwinian assumptions. Here are some basic issues:
- The mutation rate is not really known, it is inferred
- I have a paper coming out in the Fall CRSQ on "future fitness" explaining some possible reasons nearly-neutral mutations may be occurring for the benefit of the population
- The rate of beneficial mutations has not been empirically calculated. One reason evolutionists assume it is so high is that otherwise they have to think about directed mutagenesis
- Directed mutagenesis throws kinks into most of the areas, and their effects are not modeled by the software
Okay, so I probably need to explain directed mutagenesis a little more and why it impacts their model.
Historically, evolutionary thought has thought of mutations happening essentially haphazardly - that is, without any particular constraint (except perhaps incidental ones) on which DNA bases get modified. However, what if that assumption is wrong? There are several possibilities:
- Instead of mutations compounding, many of them are instead cycling.
- This means that the fixation of a deleterious gene isn't necessarily permanent. If it is cycling, it may later be replaced again with the original version.
- If the mutations are cycling, then perhaps they are all beneficial in some circumstance. Therefore, perhaps at the population bottleneck 50% (just to pull a random number out of the hat) of them switch from deleterious to beneficial.
This seems, to me, to undermine the concept of Genetic Entropy, or at least greatly reduce its effect. I don't doubt that some mutations are haphazard, I just wonder if, however, much or most of what we see are instead directed/cyclical mutations. There certainly is a lot of evidence for this (we've covered this and its mechanisms elsewhere on the blog before - if I get bored I might update this post with links - some examples include SSRs, transposable elements, and Barbara Wright's single-stranded conformation transcriptional mechanism). Genetic Entropy probably is happening, but probably nowhere near the rate that Sanford's team supposes.
I just looked at my notes again and realized I left off another source of genetic variation which is at odds with Sanford's model - Virus-borne mutations. These would affect the population in a way which is not birth-dependent. If viruses are designed to transfer new genes, reconstituted genes, or any other sort of beneficial change, then this could be a source of beneficial mutations which don't rely on the birthing bottleneck. For those of you who think that viruses are bad, the fact is that we simply don't know what most of them do. The only ones that we are really sure we know what they do are the disease-causing ones, but they are a minority.
Anyway, it would really be interesting to see the application reworked with some of these concepts in mind.