Biology Reference
In-Depth Information
lar in size and shape that the concept of atomic “homology” is mean-
ingless. Therefore, many molecular algorithms are designed for use
with unlabeled landmarks, which present a far more imposing compu-
tational problem. Because we are restricting our molecular
applications to comparisons of similar molecules, where labeled atom-
to-atom correspondences seem reasonable, our algorithm is much
simpler to implement.
7.3 Detection of phylogenetic signals
Within the past two decades, biologists have reached a consensus
regarding the importance of a phylogenetic (historical) perspective in
the analysis of comparative data. This consensus holds that a phyloge-
netic framework is essential for addressing many questions about
evolution (Coddington, 1988; Harvey and Pagel, 1991; Stearns, 1992;
Rose and Lauder, 1996). While there are a number of methods avail-
able for studying univariate data, there are currently few published
methods that describe how landmark coordinate data should be ana-
lyzed in a phylogenetic context. Part of our current research is focused
on the development of such methods.
As a first step in developing new methods, we consider a very basic
question: can we detect a phylogenetic signal in morphometric data?
When we say that there is some signal in morphometric data (whether
phylogenetic or not), we mean that similarities in form, growth, or
shape are nonrandomly distributed across the clade. In addition, if the
signal is a phylogenetic one, taxa that are closely-related will tend to
be more similar to each other than they will be to more distant rela-
tives. Put another way, we can say that morphometric data have a
phylogenetic signal if our classifications based on morphometric simi-
larities reflect genealogical relationships. To minimize the chance of
making circular arguments, we use estimates of phylogeny that are
based on data other than morphometrics (e.g., molecular sequence
data).
Before proceeding further, we might ask why phylogenetic signals
are interesting. If a phylogenetic signal does exist, it means that
descendant taxa tend to inherit aspects of their form from their ances-
tors. Furthermore, if there are aspects of form that all members of a
clade have inherited from their common ancestor, those aspects can be
considered synapomorphic (= shared derived) and, therefore, homolo-
gous. In that case, we might develop hypotheses about whether those
Search WWH ::




Custom Search