genetic vs morphological trace of phylogeny

Mon Apr 12 18:05:00 CDT 2004

In a message dated 4/7/2004 2:18:07 PM Pacific Standard Time,
dsikes at UCALGARY.CA writes:

> DNA sequence data can
> be converted to distances but they normally are not - each site is equal to
> a character of 4 states.

Sorry about the slow response, Derek, but if I stop to think,TAXACOM gets
away quickly. If, as you say, distances are not normally used in DNA studies and
instead bases are used as characters along with parsimony then I agree it's
cladists. I don't expect a sigh of relief from the molecular folks, but I did
want to respond to a later contributor who wondered if learning takes place on
TAXACOM. Although I think he meant coming around to his point of view which he
sees as learning.

Although it is a bit of an excuse for being slow to come to this conclusion,
I see a problem in the way most molecular phylogenies are presented. In those
that I've read there is a cladogram with bootstrap values and branch lengths.
No where is printed a data matrix showing sequences, nor apormorphy lists, nor
for that matter diagnoses of clades such as: related by GGTAAT in positions
8,25,46,53,54, and 89, respectively, and distinguished by AT in positions 39,
and 90 respectively. Without that information it's pretty hard to check the
work and therefore not unreasonable to conclude that the cladogram was based on
distances. After all branch lengths are what were chosen to be presented, not
apomorphic characters. Saying the authors will provide the information on
request in not a satisfactory solution and adds to Robin's question about holotypes
from molecular studies.

Since DNA is cladistic, I was also in error about the need for sequences to
have clock-like or regular mutation rates. Like morphological studies, sister
clades can diverge at different rates. Mutation rates are, as I think Pierre
said, a measure of phenetic differences between clades. Here the phenetic
information is added to a cladogram and not used to make the it. This would also
provide a test for sequences thought to have regular mutation rates. Again I
think Pierre said that, but I hate to put words in his mouth.

However, in some studies phenetic differences are used as a proxy for time
and there, I would assume, the sequences used would have to have regular
mutation rates. In these kinds of if sister clades had different mutation rates it
would mean that they split at the same time but arrived at the present at
different times.

Herb Jacobson

More information about the Taxacom mailing list