Alleles and phylogenetic results

[Karl Magnacca]

On 24 Jan 2006 at 22:29, Ken Kinman wrote:
>       However, what if different gene sequences deposited in the NCBI
> database are different alleles, are NOT recognized as such, and are
> significantly different?  Does this not have the potential to lead to
> false conclusions, especially when taxon sampling is still pretty
> sparse?

It does if the alleles are persistent through the tree (i.e. they
originate before the last speciation events).  In that case the gene is
effectively useless; you're basically talking about two paralogous genes
since the different versions have their own histories.  Still, I would
think that such a situation would be relatively rare; alleles can screw
up recent phylogenies through ancestral polymorphism, but would probably
be short-lived.  I assume you're talking about a real case though?
Where someone sequenced one allele from some species, and another from
the rest?

> I wonder how widespread this problem could be.

I suspect it's much more than realized.  A situation like this drove me
up the wall in my research on Hawaiian bees.  About a quarter of my
species had multiple haplotypes (this was in mitochondrial DNA to boot)
of up to 4% divergence within an individual.  No pseudogenes or other
wacky stuff (mostly synonymous A-G and C-T changes), and fortunately
they all appeared to be separate after diverging from the sister species
so my analysis ultimately wasn't seriously affected.  But the scarier
thing I found was that after I cloned them to get clean sequences, I
found bases that differed from ones that were unambiguous in the
original sequence (which was highly polymorphic).  So there was a lot
more variation than there appeared at first sight, and I suspect that a
lot of sequences out there that appear completely clean are hiding
polymorphisms.

Karl
=====================
Karl Magnacca, USGS-BRD, 808-985-6076
PO Box 11, Hawaii Natl. Park, HI 96718
"Democracy used to be a good thing, but now it has
gotten into the wrong hands."   --Sen. Jesse Helms