chloroplast and other genes (was Lucy in Newsweek)

[Ken Kinman]

Rob Smissen wrote:
Just to be clear, I have no reservations about chloroplast DNA being phylogenetically informative. My reservation is that it is informative about the phylogeny of chloroplasts, not their sexually reproducting hosts.
**********************************************************
Dear All,
      I would like to clarify my viewpoint as well.  In most cases, the phylogeny of chloroplasts will be congruent with their hosts.  Thus chloroplast genes did provide us valuable information in the 1990's.  My concern is that their phylogenetic signal is not particularly strong and the overreliance on chloroplasts genes could have been misleading in various cases where genetic "noise" would have made it relatively unreliable (the data thus being pushed beyond its limits).

      A better tool in my view is various nuclear genes that have a stronger phylogenetic signal, and therefore less likely to get swamped by noise.  Whole genomes will no doubt give us many different genes of this kind, and high congruency among them will give us very high confidence in those shared phylogenies (and with strong morphological congruency we would have "icing on the cake").  In those taxa where genomes provide less convincing congruencies, the morphological evidence will be more important (more than just the icing).   The big problem with prokaryotes is that we usually have both smaller genomes and usually less morphological evidence as well (and a higher level of horizontal transfer, although probably not as much as Woese might have you believe).  So genomes won't be a panacea, but will make things a whole lot easier and greatly boost our confidence levels.  And as Pierre Deleporte indicated, we have to eliminate characters that are highly vulnerable to homoplasy (or at least give them a much lower "weight" in our analyses).
              ----- Cheers,
                           Ken Kinman