[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Subject Index][Author Index]

Re: dating with dna

I received this message 10-31-95 from Bonnie Blackwell:
>I would like to challenge the dna researchers on this list who feel that 
>they can date things like the splitting of taxonomic lines with dna.
>You have not carefully examined your underlying assumptions when you make
>such statements as "dinos split from mammals 230 Ma".  That is a statement
>which can NOT be defended scientifically unless you have the dna from the
>animal where the split actually occurs.  Let me explain.
>You are erroneously assuming that all dna changes at a constant rate,
>which the fossil record clearly demonstrates is not reasonable. 
>why should change in dna occur at a static rate.  would it not be reasonable
>to assume that some changes morphology, habits, and hence the underlying
>biochemistry of the animal would require numerous changes rapidly, rather
>than at a set rate.
>for geochronologists, we never believe any date unless we can verify it by
>two independent methods, that themselves have been calibrated against
>verifiable methods.  chemically based dating methods (i.e., amino acid
>racemization, methods used for desert varnish, obsidian hydration, etc.)
>are notoriously dependent upon geochemical conditions.  similarly dna must
>depend on biochemical conditions in the body.  i'll believe dna dates
>like that when they have been corroborated by recognized geochronological
>Bonnie Blackwell,                              bonn@qcvaxa.acc.qc.edu
>Dept of Geology,                                (718) 997-3332
>Queens College, City University of New York,    fax:  997-3349
>Flushing, NY 11367-1597


Wow!  You are erroneously assuming that I am looking at DNA that does not
change at a constant rate.  Have you ever heard of molecular clocks? Clock
genes? Obviously not.  This is why, in a former message on this list, I
suggested that everyone in this field take a few genetics courses so you are
familiar with what is going on, and what will go on.  I have not the time to
instruct you in molecular genetics, but I will ramble a bit for what it is
worth.  Also, I fear you did not read my missive this AM 10-31, regarding
the fact that morphology only represents less than 2% of any organisms
genome, so please don't try to compare your rates of morphological change
with genetic change when you know nothing about the genes that code for
morphological change. They are a member of quite conserved group of genes,
some of which are called Homeobox genes. I would refer you to an article by
Sean Carrol, entitled Homeotic genes and the evolution of arthropods and
chordates, Nature, Vol. 376, August 10, 1995.

I am working with a gene called the 18S rRNA gene, which mutates at a
constant slow rate of about .8% every 50 million years(Hedges, et. al. 1990.
Mol. Biol. Evol. 7(6):607-633. Ochman and Wilson, 1987. J. Mol. Evol.
28:451-459., and lots others). Thus, by comparing any two taxa, one can
extrapolate, and indeed predict the sequence of(Felsentstein-PHYLIP
analyses) a common ancestor and also predict branching points on an
evolutionary time scale. I might also suggest that you look at an article by
Takuya Kubo and Yoh Iwasa, Evolution, 49(4), pp. 694-704. 1995, titled
'Inferring the rates of branching and extinction from molecular
phylogenies.'  The bottom line is that molecular techniques allow one to
estimate ancestral phylogeny of known taxa together with the time of
branching points in the past.  The molecular clock hypothesis proposed, and
still valid, by Zuckerkandl and Pauling, 'Molecular disease, evolution, and
genetic heterogeneity.' Pp. 189-225, in 'Horizons in Biochemistry,' edited
by M. Kasha and B. Pullman, Academic Press, NY. states that a molecular
clock is not available for morphological characters. The basis for this is
that the genes underlying your morphological characters are unknown.  The
genes underlying my data are known.

Some genes mutate inconsistently, others slow, others fast. If you pick the
right gene, you can date ancestral splits.

I have provided plenty of above references for you. Published material that
is the basis for molecular genetics. I hope this suffices, and I sincerely
hope that you read them with an open mind!

In case you are wondering why clock genes work, consider this:  ribosomal
RNA analysis indicates that ALL living organisms, and all extinct organisms,
shared a single common ancestor that probably lived out its happy days in
some hell hole pit on the bottom of the ocean in a sulfuric volcanic
scabland about 3.5 billion years ago.  Kinda makes you homesick? Doesn't it?
This means that all living organisms have evolutionary histories THAT ARE
EXACTLY THE SAME AGE. When our software analyses sequences from two or more
taxa, and a tree is constructed, the programs are actually predicting the
sequence of the last common ancestor.  Yes, there are assumtpions made. One
is that reproductively isolated clades diverge from each other, AT THE
SEQUENCE LEVEL, at equal rates. This is probably near truth.  They next
assume that roughly half of the sequence deviation that distinguishes any
two organisms was caused by mutations being fixed in each species.  This is
not true for extinct organisms. The predicted common ancestor using extant
organisms is the average of the sequence deviation between the two compared
organisms.  For an extinct organism(sequence) compared to an extant
organism, the contribution is far less than half since the extinct organism
stopped evolving a long time ago.  This is the basis for statistical proof
that one of the sequences is from a long dead organism, and this is what the
analysis shows for my sequences.  I apologize for the detailed rambling, but
they teach entire courses on this in college, and I cannot possibly hope to
instruct you without spending hours and hours. Please go read about
evolutionary genetics. I really don't have the time to continue this now.

-Chip Pretzman