If I am interpreting the original question correctly, it implies that we
should expect to see large regions in which the sequence of bases is
identical in closely related organisms, and different in more distanly
related organisms. This is not exactly what we see, however, because
different sites evolve at different rates. So, for any given linear
sequence region, closely related species might share identical bases at
sites that are "slowly" evolving, but be different at those same sites
from more distantly related species. However, those same closely related
organisms might differ from one another at other, "fast" evolving sites,
and these sites might be interspersed at various places throughout this
linear sequence. (For instance, in a protein-coding region the
third-codon positions may be essentially randomized, but the match across
second- and third-position codon sites might be identical for closely
related taxa.) (Note: "slow" and "fast" are relative terms determined by
the divergence times separating the ingroup and outgroup species.)
So, unfortunately, we don't see nice, identical regions (= collections of
sites), but mosaics, and we must look for synapomorphies at those sites
that have retained information on relationships.
Incidentally, I have been looking at mitochondrial DNA sequence in ants,
and in my group there is an insertion region that can vary from containing
NO bases to containing 152 bases. Five species from two closely related
genera (based on morphological characters) all share very large insertions
in this region (134 to 152 b.p. long), whereas the other species have much
smaller insertions (42 b.p. or less). This is not really an example of
what the original questioner asked for, because the sequences are not
identical (they look pretty much random), but would seem to be a case of a
linear sequence region as a whole providing a synapomorphy (which,
ironically, probably should be coded as a single character!).
--
Ted Schultz
ts15 at cornell.edu
