-------- Original Message --------
Subject: Re: “Quantitative description of within-group variatio n to
compare among groups
Date: Fri, 19 Jun 2009 01:37:31 -0700 (PDT)
From: Andrea Luigi CARDINI <[email protected]>
To: [email protected]
Dear Henrik,
about MorphoJ, I am sure there are more experienced users that can help.
Just a few comments (below) on your questions.
-----Messaggio originale-----
Da: morphmet <[email protected]>
Inviato il: 18 Giu 2009 - 14:38
A: morphmet <[email protected]>
-------- Original Message --------
Subject: âQuantitative description of within-group variation to compare
among groupsâ
Date: Thu, 18 Jun 2009 10:28:02 -0700 (PDT)
From: Henrik Kusche
To: [email protected]
Dear Morphometricans,
I am a diploma student and relatively new in the vast field of
geometric morphometrics and I would like to ask your advice:
I am assessing the variation in body shape of a tropical fish species
complex from 8 different lakes. My approach is based on 15 shape
encoding landmarks on the fish body and I use MorphoJ in order to
analyse my dataset which itself is based on .tps-files. The data are
corrected for allometry using residuals after the regression of size
and shape.
I am not sure which kind of control for allometry is implemented in
MorphoJ. If you're simply regressing all shapes (regardless of group)
onto centroid size or log-centroid size, I'd be careful to do it. You
should first check whether the same allometric trajectory fits all group
or you need 'group-specific' trajectories. A fairly traditional way to
check this is by using a MANCOVA (test for common slopes and homogeneity
of intercept - see TPSRegr help file for an example). If trajectories
are not the same but they're parallel, controlling for allometry might
still be an option. If they're not (significant slopes), then things are
a bit more difficult. Give a look at the book by Zelditch et al. (2004)
to learn more on this. It's either the section on regression or the one
on MANCOVA (can't remember right now, but can give you more details if
you don't find it).
A different approach is used by Mitteroecker et al. (2005 - if I
remember well). Philipp will be able to tell you more about this.
Based on a CVA, I can find a distinct morphology for the
fish of each lake. Morphologies in each lake are significantly
different from each other (whether by Procrustes or DFA). However,
some lakes have a greater (or more unevenly distributed) field of
morphospace variation within the lake than others. I see this as a
greater spread in the CVA plot. However, it is not helpful for my
further analysis to interpret the output graphs in a strictly
descriptive way (e.g. âLake A has a bigger field of variation than
Lake Bâ.
If you mean that you would like to compare the amount of variance of
different groups, I guess that's a kind of disparity analysis. Again,
you'll find an excellent introduction on this in Zelditch's book. A
simple example is in one (or more) of my own papers:
Franklin D., Cardini A., Oxnard C. E. - A Geometric Morphometric Study
of Population Variation in Indigenous sub-Saharan African Crania.
American Journal of Human Biology, in press. (PROOFS AVAIALABLE UPON
REQUEST)
Cardini A., Elton S. - The radiation of red colobus monkeys (Primates,
Colobinae): morphological evolution in a clade of endangered African
primates. Zoological Journal of the Linnean Society, in press. (PROOFS
AVAIALABLE UPON REQUEST - probably soon also avaialable on S. Elton's
website)
Nagorsen D. W., Cardini A. - Tempo and mode of evolutionary divergence in
modern and Holocene Vancouver Island marmots (Marmota vancouverensis).
Journal of Zoological Systematics and Evolutionary Research, DOI
10.1111/j.1439-0469.2008.00503.x.
Cardini A, Thorington Jr. R. W., P. D. Polly, 2007 - Evolutionary
acceleration in the most endangered mammal of Canada: phylogenetic signal
and cranial divergence in the Vancouver Island marmot (Rodentia,
Sciuridae). Journal of Evolutionary Biology, 20: 1833-1846.
Most of my stuff should now be avaialable at:
http://www.cons-dev.org/marm/MARM/EMARM/framarm/framarm.html
LOOK FOR BIBLIOGRAPHIA MARMOTARUM, CLICK ON THE LETTER C AND LOOK FOR
"CARDINI" (p. 8-9 until March 2009)
http://hyms.fme.googlepages.com/dr.sarahelton-publications
LOOK FOR "CARDINI"
These are mostly very recent papers and I fear that they've not all been
uploaded yet.Please, if you can't find something, let me know. I'll send
it to you and ask to upload missing files asap.
I would like to explore methods/statistical operations that would
permit me to obtain COMPARABLE VALUES that refer to the amount of
within-lake variation of overall body shape per lake (preferably
related to CV1 and CV2). Using these values, I would like to compare
the respective within-lake amount of variation quantitatively with
each other in downstream statistical analyses.
Be careful to use CV1/CV2 for this kind of comparison. The shape space
in which you're working has been 'distorted' so that between to within
group variance is maximized (see Klingenberg & Monteiro, 2005, Syst.
Biol.) and overfitting might be quite a serious problem. Check carefully
the cross-validated classification to have an idea about possible issues
with overfitting. I'd tend to compare the amount of shape space occupied
by groups in the full Procrustes shape space (i.e., the space you're in
after the generalized Procrustes analysis).
Good luck!
Cheers
Andrea
I am just wondering whether I could use exported PC-scores or
CV-scores (x,y coordinates) from MorphoJ for my purpose? Unfortunately
I don´t know how I could extract from these scores the crucial
information about the amount of variation. Or is there some more
suitable statistic I should use?
I would be very glad if somebody could share his advice/experiences with me.
Thanks a lot,
Henrik
--
Replies will be sent to the list.
For more information visit http://www.morphometrics.org";
target="new">http://www.morphometrics.org
Dr. Andrea Cardini
Lecturer in Animal Biology
Museo di Paleobiologia e dell'Orto Botanico, Universitá di Modena e Reggio
Emilia
via Università 4, 41100, Modena, Italy
tel: 0039 059 2056532; fax: 0039 059 2056535
Honorary Fellow
Functional Morphology and Evolution Unit, Hull York Medical School
University of Hull, Cottingham Road, Hull, HU6 7RX, UK
University of York, Heslington, York YO10 5DD, UK
E-mail address: [email protected], [email protected],
[email protected]
http://hyms.fme.googlepages.com/drandreacardini
http://ads.ahds.ac.uk/catalogue/archive/cerco_lt_2007/overview.cfm#metadata
More on publications at:
http://www.cons-dev.org/marm/MARM/EMARM/framarm/framarm.html
LOOK FOR BIBLIOGRAPHIA MARMOTARUM, CLICK ON THE LETTER C AND LOOK FOR
"CARDINI" (p. 8-9 until March 2009)
http://hyms.fme.googlepages.com/dr.sarahelton-publications
LOOK FOR "CARDINI"
--
Replies will be sent to the list.
For more information visit http://www.morphometrics.org
