Re: Rconstructing DNA (was Re: Dino-fuzz found in amber?)

[Erik Boehm <erikboehm07@yahoo.com>]

> So let's say ... you do come up
> with "likely" _T. rex_ DNA sequence for this peptide.  What then?
> What can you do with it?  You'll never know if you're right, because
> we're never going to find _T. rex_ DNA.  Not only do I regard the
> entire exercise as methodologically flawed, but I don't understand the
> point of the exercise.

This pretty much sums up my thoughts.
Yes you can say one DNA sequence is more likely than another - but you are only 
saying it is more likely because of the phylogenetic relationships.
You didn't derive that from the amino acid sequence.

I don't understand the point.

It almost seems like a way to "launder" assumptions such that they look like 
evidence.

If you only have amino acid sequences, just stick to working with amino acid 
sequences. Trying to construct a "maximum likelihood" DNA sequence sounds like 
it could lead to trouble/ false assumption, and it doesn't produce anything 
useable.