Wednesday night's colloquium went very well (Congratulations, Art!). It was civil, academic, and very scientifically oriented – thanks in large part to Ralph Seelke's fantastic talk, which was essentially a presentation of experimental results. Just being able to frame the debate in those terms is a huge win for ID. Some scattered thoughts (I've included all the questions I should have asked, but didn't think to):

1. In the panel discussion, both sides accused the other's experiments of not accurately representing the real world: Paul Nelson described the current state of ribozyme engineering (the attempt to create self-replicating RNA) as requiring too much intelligent guidance to accurately model evolution. Likewise, Gregory Weiss accused Seelke's e-coli evolution experiments of ignoring the cooperation and interactions among organisms in the real world, which make possible evolutionary changes that would be difficult otherwise. I think the issue of whose experiments better represent the real world is going to be a crucial one.

2. Seelke stated he was unimpressed with the current state of bacterial evolution experiments, because the bacteria generally evolve quickly at first and then plateu. Does anyone know of experiments where this is not the case (i.e., in which the bacteria continue to improve gradually over the whole course of the experiment)?

3. It seems to me that Seelke presented strong evidence against the theory that gene duplication can play any sort of significant role in evolution. Here is an outline of his results:
   
   He found that e-coli could easily restore function to a mutated gene if only one mutation was needed to do so. If two mutations were required, however, functionality was not restored in his experiments (a few billion bacteria over a few thousand generations). In fact, one strain actually deleted the non-functional gene(!).
   
   The point is not that such a mutation is strictly impossible. The probability of any one mutation is about one in a billion (no problem), the probability of the right two acruing is one in a billion billion. There are populations of bacteria on the earth with the probabilistic resources to overcome this. However, they are racing against the clock because while that gene is non-functional, it will be under selective pressure for removal. Furthermore, this is just too slow. The earth isn't old enough for any higher organism with a longer life cycle to wait around a billion billion generations everytime it needs a new gene.
   
   This is a big problem, since one of the major vehicles of evolutionary change in higher organisms is supposedly gene duplication, in which an essential gene is duplicated, and the second copy is then free to evolve an additional function without killing the cell. In the event that that gene requires more than one mutatation to evolve novel function, it likely will never aquire it.
   
   Can anyone cite evidence for widespread gene duplication as a vehicle for major evolutionary change? I'd appreciate experimental results, but evidence from homology would be interesting too.

4. One of the standard responses to how evolution overcomes irreducible complexity is co-option: that is, existing structures in the cell take on new function, thereby eliminating the need to build that function from the ground up. This is a potential way around Seelke's results – natural selection doesn't need to create a novel, selectively advantageous gene all at once, because it can take an existing gene and adapt it for new function in an incremental manner. I'm not an expert, but I don't think that the evidence from homology supports cooption on the massive scale this response predicts.

5. Gregory Weiss claimed during the faculty panel that RNA is, contrary to Paul Nelson's assertions, a pretty stable molecule. Can anyone point me to resources that verify/falsify this? Specifically I'd like to know how long it lasts in solution, its melting temperature, etc. compared to the same figures for DNA.

6. Paul Nelson's sparring with Gregory Weiss over self-replicating RNA was really interesting. Last time I checked, we had only succeeded in creating RNA that could duplicate a short region of itself. Apparently the field is doing much better than that, though I wasn't able to tell from the comments exactly what researchers can and can't do

7. Paul Nelson made some remarks on the problems of prebiotic RNA synthesis that bear noting. Besides the fact that ribozyme engineering assumes that RNA already exists (because there is no convincing mechanism for pre-biotic RNA synthesis), there is the question of concentration. The fundamental unit of life is the cell because maintaining high concentrations of particular substances is critical for self-replication and every other life process. The problem for origin-of-life research is not just to create self-replicating RNA but to explain how it could have been present in sufficient concentration to actually do any replicating.