-------- Original Message --------
Subject: tpsRegr and semilandmarks
Date: Wed, 3 Mar 2010 13:00:33 -0500
From: Ben Wasserman <[email protected]>
To: [email protected]

Hello,

I am investigating the relationship between geometric morphometrics
and feeding performance in two species of fish.  I have completed
separate partial warps analysis of a lateral photo of the fish's
entire body and a dorsal photo of the fishes' head (after clearing and
staining the head) using tpsRelw.

Interested in using PLS to compare shape variation to feeding
performance metrics, I then concatenated the weight matrices of the
two analyses (including uniform components) and standardized all of
the warps to mean = 0 and sd = 1 (while Rohlf and Corti 2000 suggest
it is unnecessary for a single set of warps, it appeared to be
necessary for both set of warps combined).  I computed the PLS in
MATLAB and now am attempting to visualize the results.  I'd like to
use tpsRegr to visualize the differences in shape by regressing the
PLS scores onto each set of shape statistics separately (head and
body).

tpsRegr does not seem to accommodate sliding semi-landmarks, which I
used when analyzing the dorsal view of the fish.  Am I correct in
assuming this will lead to tpsRegr calculating slightly different
partial warps than I did in tpsRelw?  If that's the case, then I think
it's incorrect to use tpsRegr like I had wanted, since the PLS scores
are calculated based on the tpsRelw version of the warps.  Is there
another program with similar capabilities that does accommodate
sliding semi-landmarks?

Thank you very much for your help!
Ben Wasserman
Research Technician
Duke University Marine Lab
[email protected]



--
For now, new message AND replies should be sent to:

[email protected]

/* Replies will be sent to the list. */
For more information visit http://www.morphometrics.org

Reply via email to