Turning around

[B.J.Tindall]

Dear Richard,
In an earlier post you indicated that no-one had replied to some of your
comments. I confess that at the time I checked out your web link and simply
came to the conclusion that you seem to be "on the ball". Since Cornelia
and I, both working in an area where gene and protein sequences are
considered to be central elements, seemed to have started this thread, I
think I should comment further.

The introduction of sequence work can be traced back via a series of
publications which date back as far as the early 1950s. Gene sequences were
not used simply because the structure of DNA was unknowm, nor how it coded
for the amino acids (far too many people tend to forget that). One key
publication was by Zuckerkandl and Pauling (1965) which clearly stated the
potential of (again protein) sequences for investigating evolution.
However, I can't recall them ever saying you can't use fossil data. They
also draw attention to possible problems. The original paper is worthwhile
reading and evaluating in a more recent context.

Another problem which arises is the problem of "linearity". This surfaces
in the misuse of the term "homology". As far back as the late 1960s
Margoliash, Fitch and Dickerson (working in those days on cytochrome c
sequences) clearly indicate that homology is not = similarity and that
there is a huge difference between linear comparisons of sequences and
looking at the three dimensionality of the sequence (= structure and
function). While a cytochrome c sequence may not be directly linked to the
form of the mammalian skeleton, cytochrome c has a very clear "structure
and function" (i.e. morphology at a molcular level). Factors such as codon
usage, amino acid substitution etc all playing different roles in how the
gene sequence can change while staying within the "rules" imposed by
structure and function - your post below. The work on the structure of the
ribosome in recent years has also emphasised this element - at least to the
(molecular) structural people.

The advent of the 16S rDNA in the txaonomy of prokaryotes was not a "single
parameter event". Ralph Wolfe, who supplied organisms to C.R.Woese, has
clearly documented this in a recent article in the reference work "The
Prokaryotes". Cell wall structures, cellular lipids, biochemical
peculiarities all contributing to establishing the usefulness of the 16S
rDNA. That molecuar methods have their place in modern systematics is
fairly evident, but in context please.

Brian

At 15:13 21.02.06 -0600, Richard.Zander at MOBOT.ORG wrote:
>The whole point of molecular systematics is that it is hoped that
>differences in DNA usually or mostly do NOT reflect evolution directly,
>because if they did, we would have the same problem with convergence as is
>the case with morphology and molecular analysis would then not be a second,
>independent test. Molecular traits are, it is hoped, not connected to
>anything in morphology. They just accumulate with each event of speciation,
>and the ones shared the most are the most ancestral. The logic is great, the
>execution so far problematic, possibly because of the inclusion of many
>apparently noncoding or codon-synonymous sequences that are actually subject
>to selection pressures.