Haplotype blocks and the punctuated equilibrium

I've just "come back" (yes, it was in Edinburgh) from PopGroup 2005, where Gil McVean presented results indicating that recombination happens mostly just before and after genes, but rarely within genes.

Nick Barton commented to the effect that this had implications for theories about the evolution of recombination; specifically (if I may extend the thought), it is clear that clonal interference will act within genes, so that only one nucleotide substitution can become fixed "at a time" (where "at a time" assumes that the time scale of fixation of alleles is small relative to the timescale over which mutation is likely to produce a second mutation within the haplotype block; essentially, this assumes a reasonably small population, somewhat less probably than typical microcosm populations).

The theory of punctuated equilibrium posits that there are brief periods in the evolution of a species where the equilibrium (eqm) of expression and enzyme activity levels is broken by usually a single nucleotide substitution, which triggers a cascade of substitutions by selection at other loci, until the system returns close enough to an eqm. The recombination data would suggest that reattainment of equilibrium will take much longer if several simultaneous changes in the coding sequence of a single gene are required. It would seem to me that this makes the punctuated equilibrium a less plausible explanation than it previously was.