genetic vs morphological trace of phylogeny

[Richard Jensen]

John Grehan wrote:

>
>
> On a related item posted earlier I have an off list response for general
> interest:
>
>  >>>I thought the goal of a cladistic analysis using parsimony was to
>  >>>identify the tree with the minimum amount of homoplasy. If this is not
>  >>>the way you reconstruct your phylogenetic trees I guess all the previous
>  >>>discussions have been a waste of time.
>
> Wrong! Parsimony finds the cladogram with the least number of steps so it
> can be compared with other 'most parsimonious' cladograms. This is why it
> is referred to as an Occcam's razor. It has nothing to do with the least
> amount of homoplasy. A cladogram with 100 steps can be less homoplastic than
> one with 99 steps.

How can a tree that has one extra step (parallelism or reversal) be less
homoplasious that the one that is one step shorter? Maximum parsimony is based on
finding the shortest tree for the data at hand.  Thus, its goal is keep homoplasy to
a minimum, i.e., find the tree with the least amount of homoplasy for that set of
data.  I'm sure that's what the offlist responder had in mind.  Of course,
characters cannot be declared homoplasious until a tree is chosen (unless one has
performed something akin to clique analysis, which can determine which characters
are likely to be homoplasious without reference to a tree).

Cheers,

Dick

--
Richard J. Jensen              | tel: 574-284-4674
Department of Biology      | fax: 574-284-4716
Saint Mary's College         | e-mail: rjensen at saintmarys.edu
Notre Dame, IN 46556    | http://www.saintmarys.edu/~rjensen